Crime as a Social Cost of Poverty and Inequality: A Review Focusing on Developing Countries

Crime as a Social Cost of Poverty
and Inequality: A Review Focusing
on Developing Countries

E Bourguignon

that criminality is not a purely structural characteristic of society which
can change only very slowly and almost ineluctably with the process of
economic, social and cultural development. It thus is important to examine
the possible determinants of that evolution. This is what we do in the rest
of this paper, first by reviewing existing economic theories of crime and
then examining the empirical relationship between criminality and possible
determinants of it.

2. Crime, Poverty and Inequality: What
Economic Theory Has to Say

The canonical theoretical model of the economies of crime goes back to
GaryBecker in 1968. It was first given some empirical content by Ehrlich
(1973). We briefly summarize the basic argument behind this model with
a simple general framework to be used throughout this paper. We then
discuss the implications of the canonical model and consider various
possible extensions likely to modify them⁹.

Let us assume that society is divided into three classes: the poor (p),
the middle (m) and the rich (r), with resources w, say wealth, such that
∞_p < w_m < tc^,_r. Let also n_m and n_r be the demographic weights of these
three classes in society. Assume that the utility function of wealth is
logarithmic and let the crime activity be represented in the following
simple manner. Crime pays a benefit equal to x with probability (1 - q)
and -F with probability q-. Thus, q is the probability of being caught, in
which case a fine F is due. We do not specify how much it is, but assuming
that it is proportional to wealth, гг?., makes things simpler. Criminal activity
is an all or nothing decision. In each class, an individual i with wealth wi
will opt for criminal activity if:

(I - q ) ∙ Log (wⁱ + x) + q ■ Log (u`. - F) > Log τr. + h. (1 )

where h. is a parameter describing the degree of ‘honesty⁷ of this individual.
It is assumed that this variable is independent of the level of income and

For a more systematic review of theoretical models of crime, see Bourguignon (1998).

70