Drug markets and COVID-19: A spatiotemporal study of drug offence detection rates in Brisbane, Australia

Abstract

In many parts of the world, the social mobility restrictions and stay-at-home orders introduced during the early months of the COVID-19 pandemic have been associated with significant reductions in crime. However, contrary to this general finding, illicit drug offence detections increased significantly. In this study, we explore the geographical distribution of the increase in Queensland, Australia, using spatiotemporal generalised additive model (GAM) to identify locations in the Local Government Area (LGA) of Brisbane where drug offence detection rates were unusually high during the three months of the COVID-19 lockdown (April-June 2020). Contrary to expectation, we find that the increase in drug offence detection rates appears to have been modest in most places, but widespread and diffuse throughout the city. We conclude that drug offence detections are most likely to have increased incidentally, probably as a consequence of general street policing initiatives which saw an increase in the visibility and vulnerability of drug user communities. We do, however, identify several locations in Brisbane where the drug offence detection rate exceeded the prediction by a considerable margin (in one case, more than double the upper limit of the prediction). We argue that in these locations the increase was likely the result of some spatial displacement of inner-city drug markets coupled with a series of targeted policing activities. Further research is needed to clarify the true mechanism of change in these locations.

Keywords: COVID-19; Drug offence detections; Generalised additive model; Pandemic; Queensland.

Fig. 1

Monthly drug detection rate (per 100,000), Brisbane LGA (April 2016 - June 2020). Source: Queensland Government Open Data Portal - Computer file.

Fig. 2

Visual representation of a GAM estimated geospatial thin-plate spline / tensor product (Drug Detections, Brisbane LGA). Source: Queensland Government Open Data Portal [computer file].

Fig. 3

Locations where drug detections were significantly higher (at p<0.05) than predicted for at least one month (April-June 2020).Source: Queensland Government Open Data Portal [computer file].Note: Areas in grey are those where over three months there was an overall ‘excess’ of drug offence detections. Prediction intervals for significance testing were adjusted using Bonferroni Correction.

Fig. 4

Number of months (of three) where drug detections were significantly higher (at p<0.05) than predicted (April-June 2020). Source: Queensland Government Open Data Portal [computer file] Note: Prediction intervals for significance testing were adjusted using Bonferroni Correction.

Fig. 5

Relative increase in drug offence detections (excess offences as a percentage of the upper 95% prediction interval, April-June 2020). Source: Queensland Government Open Data Portal [computer file] Note: Prediction intervals for significance testing were adjusted using Bonferroni Correction.

Fig. 6

Cross-classification of drug offence detections (number of months by relative increase, April-June 2020). Source: Queensland Government Open Data Portal [computer file] Note: Prediction intervals for significance testing were adjusted using Bonferroni Correction.