What motivations should govern the choice of a profiling policy in law enforcement
where decisions to search for evidence of crime – say drug sales or terrorism – may
vary with observable characteristics of the people at risk of being searched? That is
the question explored by Professor Charles Manski of Northwestern University,
writing in the November 2006 issue of the Economic Journal .
Policies that make search rates vary with personal attributes such as race and ethnicity
are variously defended as essential to effective law enforcement and denounced as
unfair to classes of persons subjected to relatively high search rates.
Choice of a profiling policy should depend on (a) what society wants to accomplish
and (b) policy-makers know about the effectiveness of alternative strategies. Professor
Manski's analysis clarifies the interplay of these forces.
Consider what society wants to accomplish. One possibility, assumed in previous
research by John Knowles, Nicola Persico and Petra Todd (KPT), is that the police
aim to maximise the ‘hit rate' of search (the chance of detecting a crime) minus the
cost of performing searches.
In contrast, Manski supposes that society wants to minimise a ‘social cost function'
with three components: the social harm caused by completed offences; the cost of
punishing offenders who are apprehended; and the cost of performing searches.
He shows that the difference between these objectives is consequential. KPT
concluded that, in the absence of discriminatory intent, police should choose a search
rule that equalises hit rates across racial and ethnic groups.
Manski finds that equalisation of hit rates is not generally optimal if the objective is to
minimise the net social cost of crime. Indeed, there are circumstances in which it is
optimal to search a less crime-prone group and not to search a more crime-prone
group. This may occur if members of the former group are deterrable and those in the
latter group are not.
Consider what policy-makers know about the effectiveness of alternative search
strategies. The consequences of candidate search rules depend on the extent to which
search deters crime. Deterrence is expressed through the ‘ offence function' , which
describes how the offence rate of persons with given characteristics varies with the
search rate applied to these persons.
Previous economic studies of profiling have assumed that the police have full
knowledge of the offence function. This assumption seems unrealistic, so Manski
studies how a social planner with only partial knowledge of the offence function
might reasonably choose a search strategy.
To demonstrate general ideas, Manski supposes that the planner observes the offence
rates of a study population whose search rule has previously been chosen. He knows
that the offence rate weakly decreases as the search rate increases, but he does not
know the magnitude of the deterrent effect of search.
In this setting, Manski first shows how the planner can eliminate some search rules
that are definitely inferior. He then examines some ways to choose reasonably among
the strategies that remain.
ENDS
Notes for editors : ‘Search Profiling with Partial Knowledge of Deterrence' by
Charles Manski is published in the November 2006 issue of the Economic Journal .
Charles Manski is at Northwestern University.
For further information : contact Charles Manski on +1-847-491-8223 (email:
cfmanski@northwestern.edu); or Romesh Vaitilingam on 07768-661095 (email:
romesh@compuserve.com).