Crime as a Social Cost of Poverty
and Inequality; A Review Focusing
on Developing Countries
R Bourguignon
on the explanatory variables that are introduced. The core independent
variables are GNP per capita, the Gini index for the distribution of income,
average education, urbanization rate, and variables controlling for the
importance of drug consumption. Among them, it is noteworthy that the
only variable more or less systematically significant turns out to be the
Gini index with, as expected, a positive influence on crime. Moreover, this
effect is sizable. All other things being equal, a 5 percentage point change
in the Gini index, which corresponds very roughly to the increase in
household income inequality observed during the 1980s in the us and in
the UK, would produce on average an increase of approximately 15 per
cent in the homicide rate, and two or three times this figure for robberies.
However, it is worrisome that, in the case of homicides, the corresponding
coefficient becomes insignificant when one controls for regions, and, in
particular, when a dummy variable for Latin America is used as an
explanatory variable. In view of the regional orders of magnitude of crime
rates reviewed above, this is not really surprising. This result suggests
that the significance of inequality’ as a determinant of crime in a cross-
section of countries may be due to unobserved factors simultaneously
affecting inequality and crime rather than to some causal relationship
between these two variables. Results obtained with robbery’ rates are more
robust. There, the coefficient of the Gini index remains significant even
when dummy variables controlling for regions or other groupings of
countries are introduced. This means that inequality appears to be
significantly associated with the crime rate within these various groups of
countries rather than mostly across them. Somehow, this is reassuring
since it fits the intuition that the economic determinants of crime are
likely to be stronger for property’ than other crimes.
Other variables do not come out significantly. This is not too surprising
for GNP per capita since most of the economic explanation of crime
somehow refers to relative rather than absolute income factors. It is less
expected that the average educational level of the population at working
age, drug consumption and the urbanization rate all turn out to be
insignificant. Measurement errors may’ affect the first tw’O variables. The
average level of education should refer to younger generations rather than
to the whole population, which may’ make a big difference in developing
countries. Drug consumption is proxied by the drug possession crime rate,
which most likely is badly recorded or a bad approximation in some
countries. The urbanization rate docs not have these problems.
Interestingly enough, it is positive and not far from statistical significance
86