International Journal on Criminology Volume 3, Number 1, Spring 2015 | Page 39

International Journal on Criminology
They refined their analysis without establishing causality. The impact of the
centralized door locking is still presented as a hypothesis—“This is consistent with central
locking as the cause of the decline because better locks would reduce door-forcing more
than window-breaking” (Farrell et al. 2008, 18)—and they state that “we hope to undertake
this further step of analysis” (Farrell et al. 2008, 18).
When they turned to this question again in 2011, they took advantage of the BCS
results but also the administrative data on vehicle-related thefts in Australia: “The data
drawn on for Australia are from the Comprehensive Autotheft Research System database,
held by the National Motor Vehicle Theft Reduction Council” (Farrell et al. 2011, 7).
They observed similarities in terms of the impact of security devices on temporary
thefts: “In Australia, . . . the temporary/permanent theft signature is uncannily similar to
that identified for England and Wales” (Farrell et al. 2011, 11).
Farrell, Tilley, Tseloni, and Mailley also introduced a working hypothesis that “would
suggest that different security devices impact differentially by crime type: immobilizers
reduce most the risk of theft of cars; alarms reduce most the risk of theft from cars; and
central locking will reduce the risk of both theft of and theft from cars” (Farrell et al. 2011,
15).
The BCS of 2006–2007 allowed them to calculate a “measure of protective effect”
(Farrell et al. 2011, 16) by type of security device and by type of crime—theft of a car,
theft from a car, and attempted theft. After taking precautions concerning the difficulty of
isolating the distinct effect of each security device, it was possible for them to note that “not
only are the protective effects distinguishable for each device but the variations across crime
type and device accord with theoretical expectations” (Farrell et al. 2011, 16).
For them, all this reinforced the security hypothesis, since at the end of their work of
“triangulation” based on multiple data sources, “the likelihood that the findings are spurious
. . . now appears low” (Farrell et al. 2011, 17).
This allowed them to produce other hypotheses, which they call the “debut crime
hypothesis” (Farrell et al. 2011, 18) and the “keystone crime hypothesis” (Farrell et al. 2011,
19). These follow on from the security hypothesis.
They consider that the reduction in the number of vehicle-related thefts and
burglaries resulting from an improvement in security measures prevents entry into
criminality since these crimes represent “debut crime.” In particular, in the case of car theft,
a “keystone crime,” it also prevents them from having a means of transportation to commit
other crimes.
The interest of these additional hypotheses is to describe a process that may have
had an impact on the crime drop as a whole and not only in terms of crimes related to
motor vehicles and burglaries. They are nonetheless difficult to test, like most of the many
reasons suggested to explain the crime drop.
In his “Closing the Doors” speech in 2012, Jan Van Dijk presented this work as a
promising extension of his theory of responsive securitization:
Farrell’s argument is that improved security has reduced opportunities for easy crimes
such as car theft and burglaries and that this has in turn blocked access to the first
stages of a criminal career for new would-be juvenile delinquents. . . . These hypotheses
suggest promising new directions in criminological research (Van Dijk 2012, 32).
38