Estimating the extent and nature of offending by Sudanese-born individuals in Victoria

Abstract

The involvement in crime of some young Sudanese-born Victorians has received sustained public attention in recent years. The media coverage of these occurrences has been extensive, with some outlets criticised for sensationalist reporting and prejudiced undertones. A range of views were held across the commentariat including, for example, the notion that Sudanese-Victorian criminal involvement has been overstated; that some level of justice over-representation was inevitable due to the demographics of Sudanese-born Victorians, which skew young and male (i.e. the demographic hypothesis); and that offending rates may be associated with heightened law enforcement responses following a high-profile criminal incident in March 2016 that received protracted media coverage and political commentary (i.e. the racial-profiling hypothesis). This paper sought to address these contentions by (i) examining the offending rates of both young and adult males across three cultural sub-groups (i.e. Sudanese-born, Indigenous Australian, Australian-born) across several offending categories between 2015 and 2018 and (ii) exploring the impact of a high-profile criminal incident in March 2016, on the offending rates of Sudanese-born Victorians. Offending rates were calculated using offender incident data from the Victorian Crime Statistics Agency and population estimates from the Australian Bureau of Statistics Census data. Findings indicate that Sudanese-born individuals figure prominently in both youth and adult offending categories relative to other major cultural sub-groups. Rates for ‘crimes against the person’ were especially pronounced for Sudanese-born youth and significantly higher than rates for crimes more subject to police discretion (i.e. public order offences). The ‘demographic hypothesis’ did not hold for the specified age range of 10 to 17 years. An increase in offending was observed post-March 2016 across two offending categories for Sudanese-born Victorians. Findings are contextualised within.

Keywords

South-Sudanese Sudanese youth crime youth offending youth violence

Australia is home to more than 30,000 people of Sudanese heritage (Australian Bureau of Statistics [ABS], 2016). Many re-settled in Australia between 2002 and 2007 through humanitarian programs following civil unrest in Sudan (Lucas et al., 2011). While the bulk of Sudanese Australians have become valued members of Australia’s multicultural community, many have faced a number of pre- and post-settlement challenges which have been outlined in prior research. Pre-migratory experiences including displacement, family fragmentation, severe life disruption, interrupted educational/vocational histories and exposure to violence often preceded post-settlement integration challenges characterised by financial hardship, low English language proficiency, unemployment, acculturation difficulties and unstable housing (Abdelkerim & Grace, 2012; Correa-Velez & Onsando, 2009; Hatoss & Huijser, 2010; Hebbani, Obijiofor, & Bristed, 2012, 2009; Khawaja et al., 2008; Milner & Khawaja, 2010; Savic et al., 2013; Schweitzer et al., 2006, 2007; Shepherd, 2016; Shepherd et al., 2018; Tempany, 2009). These co-existing social challenges have coincided with high rates of justice involvement, particularly among adolescent and young adult Sudanese Australians (see Shepherd et al., 2018). In the south-eastern state of Victoria, where approximately one-third of Sudanese Australians reside (ABS, 2016), high rates of offending by young Sudanese-born Australians have been reported across selected offending categories (i.e. aggravated robbery, aggravated burglary, assault, riot/affray) over the past three years (Crime Statistics Agency, 2018a; Shepherd et al., 2018). Between 2015 and 2018, South Sudanese and Sudanese-born individuals combined possessed an offender to population rate of 7, which was higher than the Australian-born rate of 1.1 (Goldsworthy, 2018). Across 2017–2018, young people with African ancestry comprised 19% of young people in custody in Victoria – the bulk of which were believed to be of Sudanese ancestry (State of Victoria, 2018). Evidence from 2017 (Armytage & Ogloff, 2017) and earlier (2011 to 2013) pointed to a growing over-representation of Sudanese-Victorian youth (over 4%) in youth custodial settings (Shepherd et al., 2014). At the adult level, Sudanese-born Australians have possessed the highest rate of incarceration nationwide since 2014, aside from Indigenous Australians (ABS, 2014). Evidence suggests that rates have increased substantially between 2014 and 2017 (ABS, 2017a).

‘Moomba’ and media scrutiny

The public discourse in relation to the rates of Sudanese offending in Victoria became a contentious and protracted phenomenon following a number of highly publicised criminal events in 2016 (Davey, 2016; Mills & Houston, 2016). The precipitating event transpired in March 2016, at the annual ‘Moomba’ festival, a well-attended public event in the state’s capital of Melbourne. At the event, a number of rival groups of young people engaged in sequences of violent activity, including rioting and affray, causing severe disruption. Most of the violence was attributed to a loosely connected network of predominantly Sudanese young people – described as belonging to the so-called ‘Apex’ gang – from South-East Melbourne (Galloway et al., 2016). Following this event, ongoing offending by Sudanese Victorians varied in its severity and for the most part, was ostensibly unconnected; yet, these activities were persistently and often sensationally covered by media outlets (Benier et al., 2018; Watkins & Sood, 2017). The number of articles from major newspaper outlets that made reference to ‘Sudanese’, ‘gangs’ or ‘Apex’ substantially increased post March 2016 (Benier et al., 2018; Watkins & Sood, 2017). One news outlet ran over 170 stories referencing African gangs in the year following the Moomba incident (Watkins & Sood, 2017). In 2018, federal politicians, including the Prime Minister of Australia publicly registered their concern with Victoria’s apparent ‘gang problem’ (Baidawi, 2018). The issue was also heavily politicised in the lead-up to Victoria’s state election in November 2018. The more violent (and occasionally less severe) instances of offending allegedly committed by Sudanese youth were often underscored to promote and substantiate both law and order and stricter immigration strategies (Alcorn, 2018; Chingaipe, 2018; Faruqi, 2018; Johnson, 2018; Karp, 2018; Shepherd, 2018; Simons, 2018a, 2018b; Wahlquist, 2018a). This ongoing coverage received criticism from several commentators and researchers who perceived the reportage to negatively characterise the Victorian Sudanese community (Budarick, 2018; Chingaipe, 2017; Majavu, 2018; Nyuon, 2018). Similar criticism was directed at media coverage back in 2007 when it covered the murder of a young South Sudanese male in Melbourne (Due, 2008; Nolan et al., 2011; Nunn, 2010). A recent survey pointed to the perceived social marginalisation and targeted police responses reportedly experienced by young Sudanese (and to a broader extent, African) community members as a result of the sustained and provocative news coverage (Benier et al., 2018).

The over-representation debate

The discord with respect to media ethics following Moomba extended to interpretations of official crime data. While it was regularly reported in the media that Sudanese-born Victorians were overrepresented in offending statistics, the same data was similarly invoked to underscore that Sudanese-Victorians are responsible for only a meagre proportion of overall crime (see ABC News, 2018; Goldsworthy, 2018). The latter response endured as a more pronounced over-representation for particular violent crimes become apparent.

For some commentators, the media coverage of Sudanese-Victorian criminal activity was perceived to be gratuitously disproportionate to the group’s contribution to overall crime rates, violent or otherwise (Budarick, 2018; Chingaipe, 2017, 2018; Nyuon, 2018; Watkins & Sood, 2017). Explanations for over-representation have also varied. The ‘demographic effect’ has been repeatedly prosecuted (see ABC News, 2018; Baidawi, 2018; Goldsworthy, 2018). This contention refers to the fact that Sudanese Victorians as a population are disproportionately young and male, features that also typically characterise those who are justice-involved. While this hypothesis is plausible, it has yet to be empirically verified. Episodes of over-policing and racial profiling were also believed to enhance justice system contact. From interviews and focus groups with South-Sudanese Victorian community members, it has been posited that media exposure of Sudanese offending post 2016 intensified law enforcement responses to Sudanese-Victorians leading to unwarranted targeting (Benier et al., 2018; Forson, 2019; Wahlquist, 2018b; Weber, 2018).

The study

The politicisation of Sudanese-Victorian offending over the past three years has often occasioned publicly inconsistent messaging, precluding a dispassionate analysis and understanding of the rates and causes of involvement in crime for this group. As such, it is our intention to explore and validate a number of hypotheses that have emerged during this sequence of events. First we aim to track the rates of Sudanese-born youth and adult offending at equally spaced quarterly data points between 2015 and 2018 in Victoria. We will compare offending rates across different offending categories with those of Indigenous Victorian, and (non-Indigenous) Australian-born Victorians. Australian-born Victorians denote general population offending rates and trends, and will serve as the mainstream marker and point of comparison. Indigenous Australians are consistently over-represented in offending data and thus provide an additional upper-range benchmark. The analysis will delineate the extent to which Sudanese-born Victorians are involved in crime (and which categories of crime) compared to the other cultural sub-groups. Furthermore, findings will potentially infer the extent to which Sudanese-born Victorians may be targeted by law enforcement by how their over-representation is dispersed across different offending categories. For example, Sudanese-born individuals may be over-represented in crime categories that include public order and lower level drug offences (where police discretion is wider and thus prone to bias), yet under-represented in violent offences (where police discretion is arguably restricted). Second, we will test the ‘demographic effect’ hypothesis by comparing the offending rates of young males aged 10–17 years cross-culturally (i.e. Sudanese-born, Indigenous, Australian-born) across several offending categories. Third, we aim to explore the impact of a high-profile criminal incident in March 2016 and the ensuing persistent media coverage, on subsequent Sudanese-born offending rates in Victoria by employing a time-series analyses across the period 2015–2018. It is anticipated that findings will offer some clarity to the contested public discourse on the extent and nature of Sudanese-Australian offending which will help inform strategies to reduce justice-involvement.

Method

Data

Victorian Crime Statistics Agency

Offending data was obtained from the Victorian Crime Statistics Agency (CSA). The CSA receives data recorded by Victoria police and releases publicly available datasets each quarter. The data obtained for the study encompassed the total number of criminal incidents for each quarter from January 2015 to December 2018, representing 16 quarters. We chose quarters rather than years, because this was the most detailed unit of time-based data available. A recorded criminal incident is a criminal event that may include multiple offences and alleged offenders and is recorded on a single date and at one location. Where there are multiple offences or charges recorded within one criminal incident, a single offence or charge is assigned to represent the most serious crime. The crime categories selected for the study represent classifications provided by the CSA, which are based on Australian and New Zealand Standard Offence Classification principles (Crime Statistics Agency, 2019b). These classifications include (i) Crimes against the person (i.e. homicide, assault, robbery); (ii) Property and deception offences (i.e. burglary, theft, property damage); (iii) drug offences (i.e. drug dealing and trafficking, drug use and possession); (iv) Public order and security offences (i.e. disorderly conduct, public nuisance); (v) Justice procedures and other offences (i.e. breaches of orders, resisting police officer). The number of offences were categorised by the birthplace of the individual charged with the offence, the Indigenous status of the individual and the age of the individual. Country of birth and Indigenous status are self-reported to police by the alleged offender.

The CSA does not provide data points for country of birth for quarters where the total number of criminal incidents is below 4 or where the number of criminal incidents for a country of birth is not in the top 10 countries of birth for crime in a particular quarter.

Australian Bureau of Statistics Census Data

The Census of Population and Housing (Census) is Australia’s largest national data collection. It is conducted every five years by the ABS and collects data on the demographics and key characteristics of individuals and their households in Australia. Population estimates for male Sudanese-born (10 to 17 years: n = 649; 18 years and over: n = 3387), Australian-born (non-Indigenous) (10 to 17 years: n = 221,540; 18 years and over: n = 1,289,936) and Indigenous Australian (10 to 17 years: n = 3926; 18 years and over: n = 13,356) individuals residing in the state of Victoria on census night were compiled for the study.

Statistical analyses

First, offending rates were calculated by dividing the number of offences for each quarter with ABS population estimates in Victoria for each ethnic group. The results were multiplied by 1000 to produce a rate per 1000 for each offence count. We chose to present rates per 1000 rather than 100,000, because the overall number of Sudanese males in Victoria was approximately 4000 at the time of writing, making the normal rate per 100,000 an unintuitive metric.

Second, differences in average offence rates between each group stratified by age and offence type were estimated. To achieve this, count models were constructed for each offence type and age group. Missing data was imputed using an estimate of 0 criminal incidents for quarters in which data was missing. Each model consisted of a dummy-coded term for ethnicity and an offset term which was the log of the Census population estimate divided by 1000. An offset term allows the model to account for different population sizes and allows for the modelling of rates with more accurate error terms than would otherwise be the case. We had intended to utilise Poisson regression models; however, these assume that the variance of the residuals is equal to their mean. This was not the case for the Poisson models calculated. Instead, evidence of overdispersion (i.e. variance was greater than the mean) was found. We therefore employed the negative binomial model in most cases. The one exception was the modelling of differences for youth in relation to public order offences. Here, the use of a negative binomial model led to problems estimating the shape parameter – theta – of the negative binomial distribution. Because the data was overdispersed, Poisson regression was also not an option. As such, we opted to utilise a quasi-Poisson model for this data, which is flexible enough to permit overdispersion. In total, 10 models were constructed and the alpha level was set at p < .05. Robust heteroskedastic and autocorrelation consistent standard errors (Andrews, 1991) were utilised for the analysis.

Last, we examined whether offending following the Moomba incident influenced the amount and direction of offending rates for Sudanese-born individuals. As with the previous set of models, missing data was imputed using an estimate of 0 criminal incidents for quarters in which data was missing. Negative binomial models were employed with an offset and terms for time, pre-/post Moomba and an interaction between time and the pre–post Moomba marker. Following the work of Bernal et al. (2017), the use of a categorical pre-post Moomba term allowed for the examination of abrupt shifts in offending following Moomba and the interaction term allowed for the examination of changes in the slope of the offending rate curve. Given that crime rates are known to be auto-correlated, we chose to explicitly model auto-correlation by including a lagged outcome variable in each of the models.

As with earlier models, problems estimating theta for one of the outcomes (justice procedures and other offences) necessitated the use of the alternative quasi-Poisson regression model.

Because the monthly time variable was a continuous variable, the choice of whether to model the variable as a linear or non-linear term had to be considered. To make this choice, we compared three different models; one in which time was linear, one in which time was modelled with a linear and quadratic term and one in which time was modelled with a linear, quadratic and cubic term. In several cases, more complex cubic models were not able to converge or estimate theta which was likely due to the limited sample size. In these cases, the model was not utilised and selection was limited to simpler quadratic and linear models. Model selection was based on likelihood ratio tests, and models were chosen based on whether there was a significant improvement in fit. Where a non-significant result was obtained, the simpler model was used.

Results

Comparison of offending rates over time

Figures 1 and 2 show the rate of offending across 16 quarters for each group from January 2015 to December 2018. The horizontal axis represents the three-month period of time, and the y axis shows the rate of offences per 1000 people for each of the groups examined.

Figure 1 shows the offences for males aged 10 to 17. For crimes against the person, Sudanese-born youth have the highest rate of offending (M = 65.7, SD = 24.1, range = 27.7–108.0, n = 16) relative to Indigenous (M = 20.5, SD = 3.4, range = 13.2–24.7) and non-Indigenous Australian-born youth (M = 2.3, SD = 0.2, range = 2.0–2.7, n = 16).

For property and deception offences, there was a marked elevation in the rate of criminal incidents for Sudanese born youth (M = 96.6, SD = 53.3, range = 38.5–205.0, n = 16) relative to Indigenous Australian (M = 58.2, SD = 9.8, range = 40.0–74.9, n = 16), and non-Indigenous Australian-born youth (M = 5.3, SD = 0.9, range = 3.6–6.7, n = 16). While the rate of criminal incidents was relatively stable for non-Indigenous Australian-born and Indigenous youth across the time period examined, the rate of criminal incidents for Sudanese-born youth was highly variable.

For drug offences, a lack of data for the Sudanese group meant that the average estimates were based on two time points (M = 7.7, SD= 2.2, range = 6.2–9.2, n = 2). Given the nature of missing data, this indicates that the estimate is biased in a way that exaggerates the rate of offending for this group. The Indigenous Australian group (M = 1.7, M = 0.5, range = 1.0–2.6, n = 16) and the non-indigenous Australian-born group (M = 0.4, SD = >0.1, range = 0.3–0.5, n = 16) had complete data.

The rate of offending for public order and security offences was also affected by missing data for Sudanese (M = 11.2, SD = 5.3, range = 6.2–21.6, n = 12) born youth. Both Australian-born groups had complete data, with Indigenous Australian youth (M = 4.9, SD = 1.1, range = 2.6–6.4, n = 16) having a higher average rate of offending over the three-year period compared to Australian-born non-indigenous youth (M = 0.6, SD = <0.1, range = 0.5–0.7, n = 16).

Finally, the rate of offending for justice procedures and other offences for youth indicated that the Sudanese-born group had the highest rate of offending (M = 12.3, SD = 5.7, range = 6.2–21.6, n = 9) followed by the Indigenous Australian group (M = 5.3, SD = 2.4, range = 2.8–9.9, n = 16), and the Australian-born non-Indigenous group (M = 0.4, SD = 0.2, range = 0.2–0.9, n = 16).

For individuals aged 18 years and above (see Figure 2), there were no missing data points. Sudanese born (M = 34.1, SD = 5.0, range = 26.7–41.3) and Indigenous Australians (M = 28.8, SD = 2.5, range = 24.9–33.0) had a higher rate of criminal incidents compared to non-Indigenous Australian-born individuals (M = 3.5, SD = 0.2, range = 3.2–3.8) for crimes against the person. Rates of offending for Indigenous Australian- and Sudanese-born males were similar across time.

For property and deception offences, Indigenous Australian adults (M = 47.3, SD = 6.7, range = 35.2–58.0) had the highest rate of offending relative to other groups. Sudanese-born adults had the next highest rate of offending (M= 29.9, SD = 5.8, range = 18.9–41.6), followed by non-Indigenous Australian-born adults (M = 5.2, SD = 0.5, range = 3.9–5.8).

A similar pattern of differences was evident for drug offences. Indigenous Australian (M = 5.3, SD = 0.8, range = 4.3–6.8) and Sudanese-born adults (M = 4.7, SD = 1.2, range = 3.0–7.4) possessed higher rates of offending across the three-year period relative to non-Indigenous Australian-born adults (M = 1.3, SD = 0.1, range = 1.1–1.5).

Sudanese-born adults (M = 33.3, SD = 6.8, range = 20.7–44.6) had the highest rate of public order and security offences over the three-year period. This was followed by Indigenous Australian adults (M = 14.2, SD = 1.9, range = 11.3–17.7), and non-Indigenous Australian-born adults (M = 1.7, SD = 0.2, range = 1.4–2.0).

For justice procedures and other offences over the three-year period, Indigenous Australians (M = 19.8, SD = 2.0, range = 16.2–23.1), possessed the highest average rate, followed closely by Sudanese-born (M = 16.3, SD = 3.1, range = 11.5–22.4), and non-Indigenous Australian-born adults (M = 2.3, SD = 0.1, range = 2.1–2.6).

Negative binomial and quasi-Poisson regression coefficients examining differences in offending rates for male youth and adults in each group by offence type are presented in Tables 1 and 2.

Table 1.

Coefficients and standard errors for models examining crime rates (10–17 years).

	Outcome
Term	Crimes against the person	Property and deception	Drugs	Public order and security^a	Justice procedures and other offences
Intercept	4.2 (0.1)***	4.6 (0.1)***	>–0.1 (0.9)	2.1 (0.2)***	1.9 (0.5)***
ATSI	–1.2 (0.1)***	–0.5 (0.1)***	0.5 (0.9)	–0.6 (0.2)**	–0.3 (0.6)
Non-ATSI	–3.4 (0.1)***	–2.9 (0.2)***	–1.0 (0.9)	–2.7 (0.2)***	–2.7 (0.5)***

Reference category = Sudanese born.

^aQuasi-Poisson model used.

*p < .05; **p < .01, ***p < .001. ATSI = Aboriginal and Torres Strait Islanders.

Table 2.

Coefficients and standard errors for models examining crime rates among adults.

	Outcome
Term	Crimes against the person	Property and deception	Drugs	Public order and security	Justice procedures and other
Intercept	3.5 (>0.1)***	3.4 (0.1)***	1.5 (0.1)***	3.5 (0.1)***	2.8 (0.1)***
ATSI	–0.2 (>0.1)***	0.5 (0.1)***	0.2 (0.1)	–0.9 (0.1)***	0.2 (0.1)***
Non-ATSI	–2.3 (>0.1)***	–1.8 (0.1)***	–1.2 (0.1)***	–3.0 (0.1)***	–2.0 (0.1)***

Reference category = Sudanese born.

*p < .05; **p < .01, ***p < .001. ATSI = Aboriginal and Torres Strait Islanders.

The values in Table 1 indicate a consistent pattern, whereby Sudanese-born youth had higher average rates of offending over the three-year period relative to the other groups examined. There was a single model that returned a different pattern – ‘drug offences’. Here Indigenous Australians had higher rates of offending relative to the Sudanese born group; however, the difference was not statistically significant. It is worth noting that across all offences except for crimes against the person, the differences between Indigenous Australian youth and Sudanese-born youth were not large.

Table 2 presents the results of the negative binomial regression models with adult offending data. In contrast to the youth models, the adult models showed that Sudanese-born adults had significantly lower average offending rates across the three-year period relative to Indigenous Australian adults for property and deception offences, drug offences and justice procedures and other offences. The differences between Sudanese-born and Indigenous Australian-born were not significant for the rate of drug offences. Non-Indigenous Australian-born adults had significantly lower estimated rates of offending relative to the other groups based across all offence categories.

Time-based comparison

Table 3 presents the results of the negative binomial regressions exploring the rate of offending by all Sudanese-born individuals.

Table 3.

F tests for negative binomial models measuring offending before and after Moomba.

	Outcome
Term	Crimes against the person	Property and deception^a	Drugs	Public order and security^b	Justice procedures and other^b
Time²	38.4***	12.2**	21.9	>0.1	1.6
Moomba	4.5*	42.7***	>0.1	2.0	4.5
Time – 1	>0.1	4.3*	2.7	1.0	0.5
Time²×Moomba	13.3**	9.9**	18.1***	3.7	0.7

^aQuadratic term utilised for time, Quasi-Poisson model used for justice procedures and other offences.

^bLinear term utilised for time.

*p < .05; **p < .01, ***p < .001.

The results provide evidence for statistically significant shifts in offending and changes in slope before and after Moomba for crimes against the person and property and deception offences. Drug offences had a statistically significant interaction term indicating a change in slope following Moomba, but the non-significant finding for the Moomba term indicated no sharp changes in offending directly after Moomba. Both Public order and security, and Justice procedures and other offences, showed no statistically significant changes in offending directly after Moomba or changes in slope.

The predicted rates of offending and associated confidence intervals for each offence category are shown in Figure 3. The dotted line represents the quarter in which Moomba occurred and the red ribbon represents the 95% confidence interval around the model predictions.

As displayed in Figure 3, the trend for offending in crimes against the person was increasing prior to Moomba and then stabilised following the quarter in which Moomba occurred. A spike in property and deception offences occurred following the Moomba incident and then a return to pre-Moomba levels after.

Discussion

The involvement in crime of some young Sudanese-born Victorians has received sustained public attention in recent years. The media coverage of these occurrences has been extensive, with some outlets criticised for sensationalist reporting and prejudiced undertones. During this time, a robust public debate arose in regards to the extent and significance of Sudanese involvement in crime in Victoria. Moreover, a number of hypotheses for the ostensible increasing justice-involvement of Sudanese young people emerged. These included a presumption that some level of justice over-representation was inevitable due to the demographics of Sudanese-born Victorians, which skew young and male (i.e. the demographic hypothesis). And that offending rates for particular crime categories (i.e. public order and security offences, drug offences) may be impacted by heightened law enforcement responses following a high-profile criminal incident in March 2016 that received protracted media coverage and political commentary (i.e. the racial-profiling hypothesis). This paper sought to address these contentions by (i) examining the rates of youth and adult offending across several offence categories for four cultural sub-groups between 2015 and 2018 and (ii) exploring the impact of a high-profile criminal incident in March 2016 and ensuing persistent media coverage on the offending rates of Sudanese-born Victorians relative to other cultural sub-groups.

Offending rates across cultural sub-groups

Between January 2015 and June 2018, young Sudanese-born Victorians (aged 10–17) committed crimes against the person at rates that were, on average, significantly higher than the other two comparison groups. Findings across ethnic groups for property and deception offences approximated those for crimes against the person. Offending rates were much lower for public order offences, drug offences and justice procedures offences across all ethnic categories; however, Sudanese-born youth generally possessed significantly higher rates for these crimes compared to the other groups. There were missing data points in several quarters for Sudanese-born youth across public order and justice procedures offences, and for most drug offence data points. Missing data points typically indicate a low number of offences (>4) for a group for a given quarter. There was a noticeable difference in the rates of offences for crimes against the person and property and deception offences for Sudanese-born Victorians compared to their offending rates for drug and public order offences. While not discounting experiences of profiling (which have been described in previous research), it appears that Sudanese-born young people are processed at a much lower rate for offences typically associated with profiling (public order/drug offences) than they are for offences that rely less on police discretion (i.e. crimes against the person).

At the adult level (ages 18 years and over), Sudanese-born Victorians again possessed rates for crimes against the person that were significantly higher than rates detected for the other ethnic groups. Yet, Sudanese-born adult rates for crimes against the person were substantially lower than their youth rates. The opposite effect was noted for Indigenous Australian and non-Indigenous Australian born groups (i.e. adults had higher rates for crimes against the person compared to youths). Indigenous Australians possessed higher rates of offending for property and deception offences compared to Sudanese-born Victorians. Both groups had significantly higher rates for these offences compared to non-Indigenous Australian-born adults. The same effect was found for the category of justice procedures and other offences. Again, drug offences were missing several data points, although the rates appeared to converge across all ethnic sub-groups. For public order and security offences, the rates for Sudanese-born adults were significantly higher than for all other sub-groups. Moreover, Sudanese adults possessed substantially higher offending rates for public order and security offences than Sudanese young people. The opposite effect was true for crimes against the person (i.e. Sudanese youth possessing higher rates compared to Sudanese adults).

The findings clearly indicate that Sudanese-born individuals figure prominently in both youth and adult crime rates relative to other major cultural sub-groups for the period January 2015 to June 2018. Rates for more severe crimes are especially pronounced for Sudanese-born youth. This aligns with recent custodial data indicating a large over-representation of African young people (State of Victoria, 2018). The ‘demographic effect’ hypothesis – which denotes that Sudanese justice-involvement is a function of their predominantly young and male population – did not hold for the specified age range of 10 to 17 years. Significant differences in offending rates were still detected despite the comparison of different ethnic groups by male sex and the specified age range, which effectively controlled for these factors. We recommend further tests of the demographic effect hypotheses with different age ranges. It is difficult to speculate whether the increased Sudanese-born adult rates for public order offences reflect over-policing given the simultaneous lower rates for Sudanese youth (who are typically more likely to congregate in public spaces). However, it was notable that public order offending rates for Sudanese-born adults matched their crimes against the person rates, while for the other two ethnic groups, public order offending rates were markedly lower than their crimes against the person offending rates. It is plausible that Sudanese-born adult individuals – particularly those between 18 and 25 years – may receive undue attention from law enforcement, inflating their likelihood of being processed for public order (and possibly other) offending. However, their higher rates of public order offending may also be a reflection of their over-involvement in such activities.

Offending pre–post Moomba incident

An increase in offending was observed in the post-Moomba period for crimes against the person and property and deception offences, for the entire Sudanese-born cohort. The spike in property and deception offending post-Moomba was particularly pronounced. Yet, trends for other offences (i.e. public order, justice procedures and drug offences) remained unchanged during this period. Moreover, the sharp increase in crimes against the person appeared to trend upwards before the Moomba incident occurred. It was noted earlier that crimes against the person encompass the more severe offences which are perhaps less subject to police discretion. Differential involvement in particular types offending for Sudanese-born young men appears to be the most parsimonious explanation for the recent high rates of justice-involvement.

A cautionary note must be made in light of these findings. Although the ostensibly high rates of Sudanese-born offending are of concern, only a small number of individuals are responsible for this trend. The overwhelming majority of Sudanese-Victorians are law-abiding. The prospect that recent patterns of offending may be more reflective of an increased involvement in crime for a small number of young men, rather than systemic policing bias or the natural consequence of a group’s demographics, should not discourage the search for solutions. Key risk factors for justice-involvement for this group (i.e. family disconnection, peer delinquency, school disengagement, unemployment, substance use, mental health concerns) have been identified in prior research (Coventry et al., 2015; Shepherd et al., 2018). A targeted community-informed response to these specific issues should be prioritised.

Findings from this research must be carefully interpreted in light of the following limitations. First, CSA data spanning four years (2015–2018) was compared to ABS census data from 2016 only. It is probable that the Victorian population count increased for all ethnic groups over the last two years of the study (2017–2018). As such, rates may be marginally inflated across groups for these two years (and conversely, marginally under-estimated for 2015 data points). It is unlikely that this limitation uniquely impacted Sudanese-born individuals. Second, CSA offending data refers to police charges only. All offending is therefore alleged. Some proportion of charges may not have proceeded to court or led to a conviction, possibly inflating offending rates. However, rates based on convictions only, would have omitted some perpetrators whose cases did not progress between charge and court for various reasons. Indeed, both offending classifications omit criminal activity that is not detected by law enforcement. Third, the Sudanese-born Victorian population is small and so rates of offending may appear magnified or susceptible to volatile fluctuation. To adjust for this, an offset variable was employed in our analyses. An offset term adjusts the regression model to account for differences in population. As a result, standard errors for offending rates in the model account for increased uncertainty in cases where estimates are based on a small population size. Fourth, the CSA data utilised in this study refer to criminal incidents. Criminal incidents can include multiple offences and multiple alleged offenders at one location on a particular date (Crime Statistics Agency, 2019a). The incident will be recorded by the most series offence. As a result, offending categories that encompass more minor offences (i.e. drug, public order and security) may be under-estimated. Last, country of birth of the alleged offender is estimated to be unknown in approximately 11.3% of cases (Crime Statistics Agency, 2018b).

A number of factors indicate, however, that offending by those with Sudanese ancestry is not fully captured by solely exploring Sudanese-born offending rates. For example, Sudanese-born offending rates do not include offending by individuals (aged 10 years and over) with Sudanese ancestry who were born outside of Sudan. Almost 20% (N = ∼1500) of people with Sudanese ancestry in Victoria (aged 10 years and over) were born outside of Sudan, either in neighbouring countries (e.g. Egypt, Kenya, Uganda) or Australia (ABS, 2017b). Additionally, CSA data does not separate South Sudanese from (North) Sudanese-born individuals. There is every indication that these two groups have differential involvement in crime (the bulk committed by South Sudanese). As such, we expanded the population denominator (from ABS census data) to include both South and North Sudanese-born individuals. With consideration to study limitations, it is reasonable to accept that (i) offending rates for Sudanese-born individuals, as calculated in this analysis are plausibly under-estimated and (ii) offending rates by Victorians with Sudanese ancestry are higher than rates estimated for Victorians born in Sudan, alone.

To re-iterate our earlier cautionary note, study findings should not be used to promote unhelpful narratives. Findings should instead prompt concerted efforts to better address specific community needs and stressors that increase the likelihood of justice-involvement.

Footnotes

Data

Data and replication code can be obtained from authors on request.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Stephane M Shepherd

References

ABC News. (2018, November 16). Fact check: Do Sudanese people account for only 1 per cent of crimes committed in Victoria? https://www.abc.net.au/news/2018-09-05/fact-check-sudanese-gangs-victoria/10187550

Abdelkerim

A. A.

Grace

(2012). Challenges to employment in newly emerging African communities in Australia: A review of the literature. Australian Social Work, 65(1), 104–119. https://doi.org/10.1080/0312407X.2011.616958

Alcorn

(2018, October 2). Law and order auction: How crime came to dominate Victoria’s election. The Guardian. https://www.theguardian.com/australia-news/2018/oct/03/law-and-order-auction-crime-victoria-election

Andrews

D. W. K.

(1991). Heteroskedasticity and autocorrelation consistent covariance matrix estimation. Econometrica, 59, 817–858.

Armytage

Ogloff

J. R. P.

(2017). Youth justice review and strategy: Meeting needs and reducing offending (Part 1). https://www.justice.vic.gov.au/justice-system/youth-justice/youth-justice-review-and-strategy-meeting-needs-and-reducing-offending

Australian Bureau of Statistics. (2014). 4517 – Prisoners in Australia, 2014. Retrieved from: http://www.abs.gov.au/ausstats/abs@.nsf/Lookup/by%20Subject/4517.0∼2014∼Main%20Features∼Country%20of%20birth∼7

Australian Bureau of Statistics. (2016). 2016 Census of Population and Housing. www.abs.gov.au

Australian Bureau of Statistics. (2017a). Prisoners in Australia, 2017. http://www.abs.gov.au/ausstats/abs@.nsf/Lookup/by%20Subject/4517.0∼2017∼Main%20Features∼Country%20of%20birth∼9

Australian Bureau of Statistics. (2017b). Census of population of housing 2016 (Findings based on ABS TableBuilder). Canberra, ACT: Australian Bureau of Statistics.

10.

Baidawi

(2018). Australia’s Prime Minister warns of ‘gang violence’ by African migrants. The New York Times. https://www.nytimes.com/2018/01/04/world/australia/gang-violence-melbourne-turnbull.html

11.

Benier

Blaustein

Johns

Maher

(2018). ‘Don’t drag me into this’: Growing up South Sudanese in Victoria after the 2016 Moomba ‘riot’. Full report. Melbourne: Centre for Multicultural Youth.

12.

Bernal

J. L.

Cummins

Gasparrini

(2017). Interrupted time series regression for the evaluation of public health interventions: A tutorial. International Journal of Epidemiology, 46(1), 348–355. https://doi.org/10.1093/ije/dyw0982

13.

Budarick

(2018, August 1). Why the media are to blame for racialising Melbourne’s ‘African gang’ problem. The Conversation. https://theconversation.com/why-the-media-are-to-blame-for-racialising-melbournes-african-gang-problem-100761

14.

Chingaipe

(2017, February 25). Race, stereotyping and Melbourne’s Apex gang. The Saturday Paper. https://www.thesaturdaypaper.com.au/news/media/2017/02/25/racestereotyping-and-melbournes-apex-gang/14879412004275

15.

Chingaipe

(2018, July 26). Instead of ‘African gangs’, lets discuss the complex causes of youth crime. The Guardian. https://www.theguardian.com/commentisfree/2018/jul/26/instead-of-african-gangs-lets-discuss-the-complex-causes-of-youth

16.

Correa-Velez

Onsando

(2009). Educational and occupational outcomes amongst African men from refugee backgrounds living in urban and regional Southeast Queensland. The Australasian Review of African Studies, 30(2), 114–127.

17.

Coventry

Dawes

Moston

Palmer

(2015). June). Sudanese refugees experiences with the Queensland criminal justice system (Report to the Criminology Research Advisory Council). Canberra: Australian Institute of Criminology.

18.

Crime Statistics Agency. (2019a). How the data is collected and processed. https://www.crimestatistics.vic.gov.au/about-the-data/how-the-data-is-collected-and-processed

19.

Crime Statistics Agency. (2019b). Offence classification. https://www.crimestatistics.vic.gov.au/about-the-data/classifications-and-victorian-map-boundaries/offence-classification

20.

Crime Statistics Agency. (2018a). Recorded criminal incidents by country of birth 2015-2018. Melbourne, Victoria: Crime Statistics Agency Database.

21.

Crime Statistics Agency. (2018b). What is country of birth information in police recorded crime statistics and what can it tell you? https://www.crimestatistics.vic.gov.au/media-centre/news/what-is-country-of-birth-information-in-police-recorded-crime-statistics-and-what

22.

Davey

(2016, March 14). Melbourne street brawl blamed on Apex gang after Moomba festival. The Guardian. https://www.theguardian.com/australia-news/2016/mar/14/melbourne-street-brawl-blamed-on-apex-gang-after-moomba-festival

23.

Due

(2008). ‘ Who are strangers?’: Absorbing Sudanese refugees into a white Australia. Australian Critical Race and Whiteness Studies Association e-Journal, 4(1), 1–13.

24.

Faruqi

(2018, January 3). The recent obsession with ‘African gangs’ betrays a sad reality about Australian politics. Junkee. https://junkee.com/african-gangs-australian-politics/141206

25.

Forson

(2019). The experiences of African Victorian young people. Findings and recommendations report. Melbourne: African Australian Community Taskforce.

26.

Galloway

Doherty

Jefferson

Buttler

(2016, March 14). Apex gang Melbourne: Inside the Dandenong group blamed for Moomba violence. Herald Sun. http://www.heraldsun.com.au/news/law-order/apex-gang-melbourne-inside-the-dandenong-groupblamed-for-moomba-violence/news-story/bad7e50883adf8a7a43e935bde3becaf

27.

Goldsworthy

(2018, November 5). Three charts on: Representation of Australian, New Zealand and Sudan born people in Victorian crime statistics. The Conversation. https://theconversation.com/three-charts-on-representation-of-australian-new-zealand-and-sudan-born-people-in-victorian-crime-statistics-101308

28.

Hatoss

Huijser

(2010). Gendered barriers to educational opportunities: Resettlement of Sudanese refugees in Australia. Gender and Education, 22(2), 147–160. https://doi.org/10.1080/09540250903560497

29.

Hebbani

Obijiofor

Bristed

(2009). Generational differences faced by Sudanese refugee women settling in Australia. Intercultural Communication Studies, 18(1), 66–82.

30.

Hebbani

A. G.

Obijiofor

Bristed

(2012). Acculturation challenges that confront Sudanese former refugees in Australia. Journal of Intercultural Communication, 28, 1–18.

31.

Johnson

(2018, January 3). ‘They don’t belong in Australian society’: Peter Dutton says violent African criminals who refuse to integrate should be deported from the country. The Daily Mail. https://www.dailymail.co.uk/news/article-5230711/Immigration-Minister-says-African-thugs-deported.html

32.

Karp

(2018, January 3). Peter Dutton says Victorians scared to go out because of ‘African gang violence’. The Guardian. https://www.theguardian.com/australia-news/2018/jan/03/peter-dutton-says-victorians-scared-to-go-out-because-of-african-gang-violence

33.

Khawaja

N. G.

White

K. M.

Schweitzer

Greenslade

J. H.

(2008). Difficulties and coping strategies of Sudanese refugees: A qualitative approach. Transcultural Psychiatry, 45(3), 489–512. https://doi.org/10.1177/1363461508094678

34.

Lucas

Jamali

Edgar

(2011). The Sudan-born in Australia: A statistical profile. The Australasian Review of African Studies, 32(2), 10–24.

35.

Majavu

(2018). The ‘African gangs’ narrative: Associating Blackness with criminality and other anti-Black racist tropes in Australia. African and Black Diaspora: An International Journal. Advance Online Publication. https://doi.org/10.1080/17528631.2018.1541958

36.

Mills

Houston

(2016). Melbourne CBD brawl: Who are the Apex gang? The Age. https://www.theage.com.au/national/victoria/melbourne-cbd-brawl-who-are-the-apex-gang-20160314-gnima2.html

37.

Milner

Khawaja

N. G.

(2010). Sudanese refugees in Australia – The impact of acculturative stress. Journal of Pacific Rim Psychology, 4(1), 19–29.

38.

Nolan

Farquharson

Politoff

Marjoribanks

(2011). Mediated multiculturalism: Newspaper representations of Sudanese migrants in Australia. Journal of Intercultural Studies, 32(6), 655–671. https://doi.org/10.1080/07256868.2011.618109

39.

Nunn

(2010). Spaces to speak: Challenging representations of Sudanese-Australians. Journal of Intercultural Studies, 31(2), 183–198. https://doi.org/10.1080/07256861003606366

40.

Nyuon

(2018, July 11). Why do the media demonise African Australians? The Age. https://www.theage.com.au/national/victoria/why-do-the-media-demonise-african-australians-20180710-p4zqoh.html

41.

Savic

Chur-Hansen

Mahmood

M. A.

Moore

(2013). Separation from family and its impact on the mental health of Sudanese refugees in Australia: A qualitative study. Australian and New Zealand Journal of Public Health, 37(4), 383–388.

42.

Schweitzer

Greenslade

Kagee

(2007). Coping and resilience in refugees from the Sudan: A narrative account. Australian and New Zealand Journal of Psychiatry, 41(3), 282–288. https://doi.org/10.1080/00048670601172780

43.

Schweitzer

Melville

Steel

Lacherez

(2006). Trauma, post-migration living difficulties, and social support as predictors of psychological adjustment in resettled Sudanese refugees. Australian and New Zealand Journal of Psychiatry, 40(2), 179–187.

44.

Shepherd

(2016). Criminal engagement and Australian culturally and linguistically diverse populations: Challenges and implications for forensic risk assessment. Psychiatry, Psychology and Law, 23(2), 256–274. https://doi.org/10.1080/13218719.2015.1053164

45.

Shepherd

(2018, February 7). Are African youth gangs in Australia a real threat? Psychology Today. https://www.psychologytoday.com/au/blog/prevention-now/201802/are-african-youth-gangs-in-australia-real-threat

46.

Shepherd

S. M.

Luebbers

Ferguson

Ogloff

J. R. P.

Dolan

(2014). The utility of the SAVRY across ethnicity in Australian young offenders. Psychology, Public Policy, and Law, 20(1), 31–45. https://doi.org/10.1037/a0033972

47.

Shepherd

S. M.

Newton

Farquharson

(2018). Pathways to offending for young Sudanese Australians. Australian & New Zealand Journal of Criminology, 51(4), 481–501. https://doi.org/10.1177/0004865817749262

48.

Simons

(2018a, May 18). Looking for trouble. Inside Story. https://insidestory.org.au/looking-for-trouble/

49.

Simons

(2018b, January 11). Melbourne police say Daily Mail provoked African ‘gang flare-up’ scuffle. The Guardian. https://www.theguardian.com/media/2018/jan/11/victoria-police-say-daily-mail-provoked-african-gang-flare-up-scuffle

50.

State of Victoria. (2018). Youth Parole Board Annual Report 2017-18. Victoria: Department of Justice and Regulation.

51.

Tempany

(2009). What research tells us about the mental health and psychosocial wellbeing of Sudanese refugees: A literature review. Transcultural Psychiatry, 46(2), 300–315. https://doi.org/10.1177/1363461509105820

52.

Wahlquist

(2018a, January 3). Is Melbourne in the grip of African crime gangs? The facts behind the lurid headlines. The Guardian. https://www.theguardian.com/australia-news/2018/jan/03/is-melbourne-in-the-grip-of-african-gangs-the-facts-behind-the-lurid-headlines

53.

Wahlquist

(2018b, January 6). ‘We’re not a gang’: the unfair stereotyping of African-Australians. The Guardian. https://www.theguardian.com/australia-news/2018/jan/06/were-not-a-gang-the-pain-of-being-african-australian

54.

Watkins

Sood

(2017, April 21). Did the Herald Sun invent the Sudanese youth Apex gang? Crikey. https://www.crikey.com.au/2017/04/21/did-the-herald-sun-invent-the-sudanese-youth-apex-gang/

55.

Weber

(2018). ‘Police are good for some people, but not for Us’: Community perspectives on young people, policing and belonging in Greater Dandenong and Casey. Australia: Monash Migration and Inclusion Centre, Monash University.