Crime and Development in the Caribbean: An Investigation of Traditional Explanatory Models

• Author: Richard R. Bennett, William P. Shields and Beth Daniels
• Pages: 401-429

401

CRIME AND DEVELOPMENT IN THE CARIBBEAN

Crime and

Development in

the Caribbean:

An Investigation

of Traditional

Explanatory

Models1

Richard R. Bennett,

William P. Shields and

Beth Daniels2

Eighteen

INTRODUCTION

For the past quarter of a century, researchers

have been intrigued by the relationship

between societal development and crime.

Their research has been driven primarily by

two theoretical perspectives. The

Durkheimian perspective, the oldest and

most extensively investigated model,

suggests a link between societal development,

anomie caused by change, and deviance

and/or criminality (c.f. Durkheim 1964,

Merton 1959). The application of this

approach to cross-national analysis of crime

is demonstrated in the works of Krohn

(1978), Messner (1982), Neuman and Berger

(1988), Bennett (1991a), and Bennett and

Shelley (1985), to name a few. The second

and more recent perspective links

opportunity or routine activities patterns

with changes in levels of criminality (c.f.

Cohen and Felson 1979, Maxfield 1987).

Although examples of this approach abound

with regard to single nation research, there

are fewer current examples in the arena of

cross-national development (c.f. Kick and

LaFree 1985, Anderson and Bennett 1996).

These two perspectives have generated

a body of cross-national research findings

that is surprisingly and consistently similar.

First, development affects crime. Second,

development affects crimes of violence and

theft differently: As development increases,

violent crimes decrease while crimes of theft

increase (c.f. Bennett 1991a, Bennett and

Shelley 1985, LaFree and Kick 1986, Krohn

402

CRIME, DELINQUENCY AND JUSTICE

1978, Wolf 1971). Although these studies inform the scholarly community about

the general relationship between development and crime, they may not distinguish

among specific nations or regions of the developing world, nor these nations’

experiences with criminality. Thus, these studies may not provide an adequate

knowledge base from which to construct informed crime control or prevention

policy.

The problem of relevancy to regional or national policy may be due to the

methodologies employed in these studies. First, the current cross-national research

divides the nations of the world for which data are readily available into two

groups: developed and developing. As one might expect, the range of values on

critical variables within each group is often greater than the range in means

between the levels of development, if only at the margins.

Second, notwithstanding reliable data and the availability of valid measures

of the criminologist’s constructs, current analysis practices cannot resolve the

problem of specificity. The aggregation of nations into groups obscures potentially

important country-specific characteristics. What emerges is a statistical average

that, in fact, may not be representative of any nation within the group. For

example, an average could include one nation where inequality is dramatically

pervasive, and another in which there is little inequality. The average would be

moderate inequality, which is found in neither. This problem is further

compounded by possible interactions or non-linearity among the variables within

the analysis model. As a consequence, the effects of inequality could be

significantly exaggerated or underestimated.

Third, these empirical studies only include variables within their analysis

models that are commonly measured and collected by leading world-wide

organisations, including the World Bank, the United Nations (UN) and the

International Labour Organisation (ILO). Although data from these sources are

helpful to understand the relationship between development and crime, they

include aggregated data which are of interest solely to the organisations’ missions.

Data of greater interest to criminal justicians, derived from theoretical approaches,

are not often collected. Direct measures of constructs as social stratification,

opportunity structure, family structure, community cohesion, informal and formal

social control, and other ‘social facts’ are not available from the international

organisations. Thus, criminal justicians who employ these archival data must use

them as crude surrogate measures of their more complex constructs. As an example,

researchers have used the measure of women in the workforce as a surrogate

measure of informal social control or guardianship (c.f. Bennett 1991b). Similarly,

urbanisation is commonly used as a surrogate measure of community cohesion

(c.f. LaFree and Kick 1986).

This paper attempts to investigate the problems in aggregation and analysis

of existing data by combining a longitudinal analysis across three seemingly similar

developing nations and comparing the effects of traditional explanatory variables