The private harms of detention: why Serco’s violence is not criminalised

ABSTRACT

Immigration detention centres in Australia have been managed by corporations for the Commonwealth government for the last 25 years. Complaints of officers’ violence, including the use of tear gas and batons on detainees, have continued throughout. In recent years, reports suggest that officers tend to use force as a first resort to manage detainees’ conduct and routinely handcuff detainees for medical and other appointments outside detention. This article is titled ‘the private harms of detention’ not because officers’ use of force in detention is always hidden, but because few officers have faced criminal prosecution. I argue that flaws in the design and operation of risk assessments unduly heighten the risks posed by detainees. By justifying a greater level of force as reasonable, risk assessments mitigate the potential for officers to be held criminally culpable, and thereby diminish understandings of the seriousness of officers’ harmful use of force. The privatised nature of detention and ongoing controversy over the extent of power private detention operators should wield make it important to explore how such actors might influence forums of accountability.

KEYWORDS:

• Accountability
• privatised immigration detention
• private security
• risk assessment tools
• use of force

Introduction

Harm is so well-documented in immigration detention in Australia that it might as well be its synonym. The slow violence of prolonged detention has resulted in suicide, self-harm, and the deterioration of mental and physical health of those incarcerated. The tall walls, razor wire fences, and CCTV have witnessed officers spraying clouds of tear gas, throwing rocks, using fists, batons and boots against adult and child detainees. The officers involved are primarily staff of private corporations, authorised as ‘officers’ by the Minister, and hold specified powers to use force under the Migration Act 1958 (Cth) (‘the Migration Act’). Yet, over the 25-year history of privatised management of mandatory immigration detention, officers’ harmful use of force in detention has largely escaped criminal culpability for assault and other offences which might otherwise be expected.

Statutory bodies with investigation responsibilities have recently expressed concern over the use of force, including handcuffs, by Serco Australia Pty Ltd (‘Serco’) officers on detainees (AHRC, 2019b; CO, 2021). Serco is the third and current corporation contracted to provide ‘immigration detention services’ since the Commonwealth privatised immigration detention in 1998 and has undertaken this role from mid-2009 (Penovic, 2014, pp. 12–13).¹ The Australian Human Rights Commission (AHRC) has found Serco officers have used force that was not reasonable, necessary, or proportionate to the risk the detainee posed to themselves or others (AHRC, 2019b, p. 82; AHRC, 2019c, p. 47). The Commonwealth Ombudsman has noted an increasing trend in officers using force as a first resort to resolve ‘non-compliant’ behaviour by detainees (CO, 2020b, p. 23), and it seems that by 2018, Serco officers routinely used restraints to escort detainees to appointments outside detention (AHRC, 2019b, p. 31).

Despite extensive documentation of the harm that private detention officers perpetuate, the significance of such harm is masked by not being subject to criminal prosecution. The criminal law’s focus on individual culpability has long been found to poorly address harm perpetrated by corporations (Colvin, 1995; Crofts, 2017; Fisse & Braithwaite, 1993), and as I will show, risk assessment practices can contribute to individual officers also not being held criminally culpable. My main argument is that flawed risk assessment practices legitimise private detention officers’ use of force and restraints on detainees and thus curtail the potential for criminal law’s operation as a forum for accountability. The privatised nature of immigration detention plays a role as well. Analysis of the design and practice of risk ratings of detainees and of the legal regulation of Serco’s use of force on detainees will show how, as scholars have theorised, private detention centres involve public and private entities in mutually constituting the applicable law (Doty & Wheatley, 2013, p. 428; Golash-Boza, 2009). This argument draws on insights from human rights and oversight inquiries in Australia between 2000 and 2023, with particular focus on the period from 2018 when Serco’s use of force and restraints for escorts outside detention became increasingly widespread.

The article starts by establishing that Serco’s use of force is harmful even though few officers are criminally prosecuted. The main part of the article is focused on the role of risk ratings in mitigating officers’ criminal culpability for assaults and other offences. It explains why the risk assessment tool in use since 2015 is flawed in design and practice and contributes to unduly inflating the risks posed by detainees. It explores how the criminal law makes such risk assessments relevant to officers’ criminal culpability. It then identifies potential incentives for Serco to impose high risk ratings for detainees. Together this analysis shows how the flawed design and operation of risk assessments serve to resolve persistent ambiguity regarding the legal boundaries of officers’ use of force in favour of private immigration detention service providers. The combination of applicable statutory, contractual, and procedural frameworks construct private detention officers’ use of force as lawful not criminal, which masks and sustains Serco’s routine violence.

The use of force as harm

A lengthy history, which pre-dates the introduction of a formal security risk assessment tool, documents harm done by officers against detainees. There is plenty of sobering evidence for instance of officers perpetrating harm against children by using force instead of other responses. For example, an Australasian Correctional Management Pty Ltd (ACM) officer struck a seven-year-old boy across the legs with a baton (HREOC, 2004a). ACM officers pinned up an eleven-year-old unaccompanied boy and hit him across the face with enough force to leave bruises on his face and welts on his neck (HREOC, 2004b, p. 339). Serco officers used an ‘enhanced escort position’ including a head-lock style position to force unaccompanied children to move from one compound to another at Christmas Island in 2014, and although the Department of Immigration and Border Protection (‘the Department’)² approved the use of force, officers did not use force as a last resort, contrary to Australia’s human rights obligations (AHRC, 2014, pp. 160–162).

From 2015, with changes to the Security Risk Assessment Tool (SRAT) which had been introduced in 2012, the cohort of detainees classed as potentially posing a high-risk or above broadened, and thus a greater number of people were subject to policy requiring the use of physical restraints during transport and escort. It is thus not surprising that one of the ‘most commonly-raised concerns by detainees’ since then has been the use of mechanical restraints, such as handcuffs, during escorts outside detention for medical and legal purposes (AHRC, 2019b, p. 31; see also CO, 2021; CO, 2022). Any use of force (including restraints) by an officer against a detainee involves the danger it will amount to ‘abuse or ill-treatment’ (CO, 2022, p. 51). Detainees and advocates have raised concerns about the use of handcuffs since at least 2008 (JSC, 2008, p. 66). However, today people are held in detention for longer periods, and thus face a longer period of exposure to being handcuffed while escorted outside detention or between facilities. In 2023, people are detained on average for seven times as long as a decade ago (DHA, 2023, p. 12). The average period in detention at 31 March 2023 was over two years (732 days), with 130 people (11.6%) having been detained for over five years (DHA, 2023, p. 12).

The harmful effect of handcuffs exceeds their actual use. The Ombudsman expressed concern that detainees forwent offsite medical appointments to avoid being handcuffed (CO, 2022, p. 55; see also AHRC, 2019b, p. 56), an experience some detainees reported was ‘humiliating’ (CO, 2022, p. 55). Detainees in highly constrained hotel alternative places of detention (APOD) ‘often declined’ to engage in physical activity, their only access to fresh air, while cuffed (CO, 2022, p. 53). The AHRC (2011, [14]) reported claims that a detainee missed a tribunal hearing because he did not want to be handcuffed.

During the period 1 October 2021 to 30 September 2022, there were about 1500 people in immigration detention facilities in Australia at any one time (DHA, 2021, p. 8; 2022, p. 4), during which time there were a total of 7405 incidents recording officers’ use of force; more than 75% were planned uses of force requiring prior risk assessment and Australian Border Force (ABF) approval (see Table 1). Although there are no children in detention according to the latest statistics available in 2023, about 152 children were detained over the four-year period covered in Table 2 (RCOA, 2023),³ making the total of 65 incidents of force on children in detention significant.

Table 1. Use of Force incidents by incident type for people in immigration detention centres (IDC) and facilities (ITA, APOD) between 1 October 2021 and 30 September 2022 (Commonwealth, 2022).

Use of force incidents	Unplanned use of force	Planned use of force	Total
Adelaide Immigration Transit Accommodation (ITA)	<5	<95	<100
Brisbane ITA	120	3024	3144
Melbourne ITA	104	497	601
North West Point IDC	57	151	206
Northern Alternative Place of Detention (APOD)	0	11	11
Perth IDC	71	97	168
Villawood IDC	1159	1255	2414
Yongah Hill IDC	307	455	762
Total	1823	5584	7405

Table 2. Type of use of force used on detainees under 18 years in immigration detention centres and facilities between 1 January 2016 and 31 December 2020 (Commonwealth, 2021).

Type of use of force used on detainees under 18	No. detainees
Planned use of mechanical restraints for escort to offsite location	50
Unplanned use of restraints for escort within or outside detention facility	6
Unplanned use of force to prevent self-harm	<5
Unplanned use of force to prevent assault	<5
Unplanned use of force to prevent injury	<5
Unplanned use of force to prevent escape	<5
Total number of detainees under 18 involved in incidents of use of force	65

The use of force seems to have become an ordinary part of the physical harms inflicted in detention. In 2018 the use of restraints for transfers and escorts outside detention had become ‘widespread and routine’ (AHRC, 2019b, p. 31). In 2020, the Ombudsman expressed concern that:

There appears to be an increasing tendency across the immigration detention network for force to be used to resolve conflict or non-compliant behaviour as the first rather than last choice and can be exercised in a manner both inconsistent with the department’s procedures and possibly without legal basis. (CO, 2020b, p. 23)

Ombudsman and AHRC investigations into multiple complaints have found that force, including restraints, has been used contrary to policy as a first resort against both adults (CO, 2021, pp. 6–7) and children (AHRC, 2019b, pp. 13, 17, 98). Most recently, the Ombudsman reported that the unplanned use of force ‘appears to be used by default for many routine offsite transport and escort facilities’ at Villawood IDC, and along with other centres, rules for the use of force were not followed (CO, 2023, p. 21). Restraints have been used on people in wheelchairs, people with low security risk ratings, and on a man with a mental illness and no history of aggression (AHRC, 2019b, p. 17). Where used, a person is restrained for the entire duration of the escort which can be most of the day and restraints may be left on while a person is trying to eat or use the toilet (AHRC, 2019a, p. 30).

The tendency for restraints to be recorded simply as ‘mechanical restraints’ or ‘MR’ has hindered understanding of the type of restraints used in each incident (CO, 2020a, p. 19). But inquiries show the following use of restraints: flexi-cuffs (800 g maximum security handcuffs, the use of which for extended trips the Ombudsman has, from 2017, repeatedly recommended against); spit hoods (for example, in the period 1 July 2021-30 June 2022 there were five reported incidents of spit hood use within immigration detention), and head protection (CO, 2020a, p. 24; 2020b, p. 22; 2023, p. 23).

Aside from the use of restraints, other recent uses of force are also concerning. Firefighting extinguishers and equipment were used in 2022 against detainees during two incidents at North West Point IDC on Christmas Island in breach of rules against the use of chemical restraints in a ‘pre-planned and systematic use of force’ (CO, 2023, p. 22). Reports that officers allegedly assaulted or used disproportionate force on detainees have persisted in oversight reviews and include allegations of rough and invasive searches (AHRC, 2019a, pp. 23, 54), and physical restraint of people on the floor despite detainee compliance with instructions (‘mandatory ground stabilisation’) (CO, 2021, p. 7). Some facilities demonstrated an ‘over-reliance’ on pat searches (CO, 2022, p. 53) and there were reports of regular room and property searches, as well as pat and wand searches such as before and after visits and meals (AHRC, 2019a, p. 54).

While the AHRC and the Ombudsman reports show the nature and extent of the harm involved in Serco’s use of force, much of what has been described in this part is regarded as the ordinary and non-criminal use of force approved by the Department to manage immigration detention. Many of the harms of detention have been licensed by policy. For example, in the early 2000s, Department policy permitted, and officers used, tear gas and water cannons on detainees to quell protests (HREOC, 2004b, pp. 300–301, 309). This article follows the approach of a diverse scholarship premised on the notion that harm itself should be the framework for study, rather than conduct legally labelled as ‘criminal’ (Canning & Tombs, 2021; Hillyard & Tombs, 2004; Leighton & Wyatt, 2021; Tombs, 2018). For detainees, the experience of harm surpasses that within the purview of criminal law and includes harm to physical and psychological health and financial harm (see for example AHRC, 2014; HREOC, 2004b). Attention to harm that may fall outside the boundaries of the criminal law and legal system is also an impetus in crimmigration (Billings, 2019; Stumpf, 2006) and border criminology literature (Aas & Bosworth, 2013; Bosworth, 2019; Weber & McCulloch, 2019). For example, asylum seekers and other detainees typically experience immigration detention as punishment akin to criminal legal system imprisonment, with many harmful effects (Peterie, 2018). This is the case even though visa cancellation, which is a common administrative precursor to detention in Australia, does not formally impose punishment (Billings & Hoang, 2019; Grewcock, 2014). Serco’s use of force is demonstrably harmful even if not criminally prosecuted.

Criminal prosecutions

There have been few criminal prosecutions of staff of private immigration detention service providers. One notable exception is the conviction of a G4S guard and a Salvation Army worker (contracted to provide welfare services) for the murder of Iranian asylum seeker Reza Barati in Australia’s detention centre on Papua New Guinea’s (PNG) Manus Island in 2014. The two staff members were convicted in PNG’s Supreme Court and sentenced to ten years imprisonment with five years suspended. However, the court heard evidence that Australian and New Zealand guards were also involved in the attack on Barati (Doherty & Davidson, 2016), yet no charges were laid by the Commonwealth Department of Public Prosecutions (CDPP). Within Australia, a Serco officer at Melbourne ITA in Broadmeadows was convicted for indecent assault of an asylum seeker detainee in 2015 (Lee, 2017). However, convictions are far outnumbered by reports of harm by private detention officers (see for example AHRC, 2019a; 2019b; 2019c).

Why do few prosecutions of private detention officers proceed? In some instances, detainees have requested the Ombudsman not pursue individual complaints, including over the inappropriate use of force and harassment, ‘largely because of the fear of retribution’ (CO, 2001, p. 26). Detainees have also explained they have not reported alleged assaults by officers due to fear it could affect their migration legal processes (Commonwealth, 2006, p. 183). There may also be delays in the ABF referring matters to police, as occurred in one case where a detainee requested that an alleged assault on him by another detainee be referred to police investigation but due to human error the incident was not referred for about 10 months (AHRC, 2023, pp. 19, 35, 37).

In circumstances where police pursue charges, the unlawful non-citizen status of the victim witness might result in removal from Australia and frustrate prosecution. For example, in 1996 the Australian Federal Police (AFP) referred allegations of Australian Protective Service (APS) officers assaulting and injuring detainees to the CDPP, but the CDPP did not prosecute because the witnesses had been removed from Australia (HREOC, 1998, pp. 96–97). At other times, the CDPP may withdraw charges for reasons that similarly arise if the victims are citizens, such as in a 2001 case in which the ACM officers who allegedly assaulted an eleven-year-old could not be located and one had moved overseas (HREOC, 2004b, p. 339). Cases of alleged abuse of power, detainees’ reluctance to report assaults by officers, and the cessation of police investigation if the victim was deported, were matters that a Parliamentary Inquiry in 2006 said required ‘concrete steps to combat criminal activity and the culture of impunity’ (Commonwealth, 2006, p. 184).

In contrast to the few prosecutions of immigration detention officers, the AHRC has made numerous findings that officers have used ‘excessive force’ against detainees. For example, the AHRC found five guards had applied force to Mr FE (pseudonym), an asylum seeker detained at Christmas Island in 2015, that was ‘unnecessary, excessive and contrary to both Serco’s Operational Safety Manual and the Department’s Detention Services Manual’ (AHRC, 2019b, p. 82). The force caused Mr FE’s tooth to be knocked loose (AHRC, 2019b, pp. 82–83). In another instance the AHRC (2019c, p. 47) found that Serco Emergency Response Team (ERT) officers used ‘excessive force’ on Mr FZ (pseudonym) on two occasions. At the time of the use of force Mr FZ was isolated in a secure room, and the AHRC considered the number of ERT officers (seven officers with up to six using force at the same time) and the way they restrained Mr FZ on the ground ‘was not reasonable, necessary or proportionate to the risk he posed to himself or others’ (AHRC, 2019c, p. 47). The use of force resulted in injuries to Mr FZ’s hand and possibly also to his ankle (AHRC, 2019c, p. 55).

While the AHRC has used terminology that is relevant to the assessment of criminal culpability, the AHRC makes findings where it is reasonably satisfied of the facts of those findings on the balance of probabilities and is not required to reach the criminal standard of proof of beyond a reasonable doubt. In both these cases, officers’ use of force was referred to the AFP, but neither resulted in charge (AHRC, 2019b, pp. 81–82; 2019c, p. 10). It has never been the case that all suspected criminal offences be prosecuted (see for example Goodrich, 1986, pp. 18–19), and it is not clear whether the purported risks posed by Mr FE and Mr FZ influenced the AFP’s decision not to charge. However, as the next two parts contend, Serco’s role in the assessment of the risks that detainees pose inform whether officers will use force and are also relevant and meaningful to officer’s criminal culpability.

Risk ratings in immigration detention

Serco holds contractual responsibility to conduct risk assessments for people in detention (AHRC, 2019a, p. 21). Serco’s risk assessments have informed everything from use of force, where a detainee will be detained, and whether a detainee will be restrained in handcuffs or a spit-hood when travelling to and from detention including to courts, tribunals, and medical appointments. For over a decade, risk assessments have utilised the Security Risk Assessment Tool (SRAT), an artefact of public-private partnership. Serco developed the first version of the SRAT following the Department’s acceptance of recommendations arising from reviews that critiqued the quality of Serco’s risk assessments (AHRC, 2019b, p. 35). Serco’s initial SRAT was approved by the Department and implemented in detention centres by the end of July 2012 (AHRC, 2019b, p. 35). In 2013, Serco recommended to the Department that risk categories be broadened; in 2015 the Department provided guidance about who should be regarded as ‘high risk’, and Serco introduced key features of their newly developed SRAT v2 in November 2015 (AHRC, 2019b, p. 36; Serco, 2015, p. 2). This part shows how key features of the design and operation of SRAT v2, referred to here as SRAT, have unduly inflated the risk ratings attached to a detainee. Flawed risk ratings serve as an unjustified basis for use of force, and shape perceptions of what amounts to a necessary, reasonable, and non-criminal use of force by officers.

Risk management tools are ostensibly intended to assist officers to mitigate risk in a uniform manner by providing a risk rating for individuals based on a range of factors. Theoretically, the SRAT is meant to assess risks such as whether detainees will escape, threaten the safety of staff and others, or self-harm (AHRC, 2019a, p. 21). However, the Griffith Criminology Institute (2019, p. 117), contracted by the ABF to independently assess the validity of the SRAT, concluded that the ‘tool is not borne out of sound scientific research’.

A large part of the problem is that SRAT’s design inaccurately ascribes risk. The Department’s full guidance to Serco in 2015 about the cohorts of detainees who should be categorised as ‘high risk’ is not available; however, it provided that the following cohorts were to be considered ‘high-risk’ at the time of entry to detention:

• any single adult without ‘physical impediments that would impair their ability to overpower or abscond from escort staff’ and

• who had either been in detention for less than 30 days (later changed to 28 days) or ‘had a criminal background involving violence or aggression, or any attempted (or actual) abscond/escape, violent or escape oriented history from detention or any form of custody’. (2019b, p. 36)

Serco argued that broadening the cohort classed as ‘high-risk’ to include almost all at entry was needed because most escapes in the first 30 days of detention during the period 1 January 2010 to 30 August 2015 were by those classed as ‘low risk’. However, the AHRC noted that escapes by low and medium risk rated detainees in their first 30 days of detention accounted for less than 0.15% of the total number of people being detained and challenged the fairness of this forming the basis for categorising all able-bodied single adults as ‘high risk’ during their first 30 days detained (AHRC, 2019b, p. 36, 37).

The 2015 changes resulted in the requirement that all ‘those for whom the risk assessment indicates that they potentially pose a high risk’ be restrained during escorted visits and scheduled travel, including all able-bodied detainees in the first 30 days of detention (the Department’s 2015 Detention Services Manual cited in AHRC, 2019b, p. 32; 2019b, p. 36, 132). There have since been changes to the policy—‘high-risk’ categorisation applies in the first 28 days only and ‘does not apply where there is sufficient knowledge/information of the detainee to inform the risk assessment’ (AHRC, 2019b, p. 38; see also DHA, 2018, p. 8). The Department has also emphasised that a detainee’s overall risk profile, which informs planned used of force, is based on individual assessment (AHRC, 2019b, p. 149). However, the AHRC concludes that the approach ‘still involves treating some people as high risk regardless of their actual level of risk’ and it remains likely that restraints will be used in escort of people who should be categorised as low-risk (AHRC, 2019b, pp. 15, 38).

The Ombudsman (2020a, p. 17) found that the SRAT design was not based on established sociological and psychological research on the likelihood of reoffending and of violent behaviour. The SRAT places disproportionate weight on criminal history. For example, a detainee with a violent criminal history is always ranked as high risk even if there has been a lengthy period since any offending and they have since undertaken rehabilitation (CO, 2020a, p. 17; see also AHRC, 2019a, p. 21). In some instances, disproportionate weight has been placed on offences committed 15 or 20 years prior (CO, 2020b, p. 42), and risk ratings have not changed despite later positive conduct (AHRC, 2019a, p. 22). The Department’s own decisions signify that many detainees pose little risk. In 2018–2019 the Department revoked over forty per cent of mandatory visa cancellations (Commonwealth, 2020b, p. 13) and thus released detainees, in accordance with directions which require decision-makers consider the risk to the Australian community should the non-citizen commit further offences (Minister for Immigration, Citizenship and Multicultural Affairs 2023, pp. 5–8).

The SRAT’s categorisation of risks also contributes to amplifying the risk ratings of some detainees. Specific convictions were linked with types of risks which can produce misleading results and ‘automatic (and at times, unwarranted) escalation’, such that ‘a person may be assessed as having links to organised crime and an associated higher risk rating without any material facts to support that rating’ (CO, 2020b, p. 17). The SRAT uses broad categories which group a wide variety of ‘risks’ together. Bad language and shouting without physical aggression is grouped with physically ‘aggressive behaviour’ under the behavioural risk indicator of ‘aggression’, with the result that detainees may be classified as ‘high-risk’ on a spurious basis (AHRC, 2019b, p. 40).

While on the one hand, the capacity for bad language to lead to a ‘high-risk’ assessment might be regarded as a classification mistake to be rectified, conceivably it is part of its disciplinary function and design. Offensive language remains a criminal offence despite multiple reports and inquiries recommending its monitoring and reform (see for example ALRC, 2017, p. 423, pp. 425–428; RCIADIC, 1991, vol. 5). Scholars have explained that ‘offensive language’ remains criminalised because it bolsters police power and authority in street order (Methven, 2020; 2018). Charges for swearing at police have long been theorised as reflecting police perceptions that swearing represents a ‘symbolic of lack of respect for authority, trouble, losing control, and [is] indicative of potential danger’ (Travis, 1983, p. 214). It is plausible that the inclusion of bad language in the broader SRAT category of ‘abusive/aggressive behaviour’ performs a similar function—bolstering Serco’s legitimacy to use force to manage perceived challenges to officers’ authority.

In any case, as the Australian Human Rights Commissioner commented, this seems especially excessive for those detained for a very long time, as ‘it might be understandable that they engage in bad language, even outside of the context of being aggressive’ (Commonwealth, 2020a, p. 19). Vogl’s and Methven’s (2020, p. 69) analysis of a Code of Behaviour imposed on asylum seekers in the community provides further insight as to why ‘bad language’ can lead to the use of force and handcuffs. The Code of Behaviour included conditions such as to refrain from ‘anti-social/disruptive behaviour’, ‘public nuisance’, and ‘causing a disturbance’ which Vogl and Methven (2020, p. 69) argued demonstrated preoccupation with ‘civilising discourses’ in which racialised others were required to ‘learn from and adopt the customs of the new host state over their own cultures’.

Serco staff operation of the SRAT also contributes to amplifying detainee risk ratings. Detainees have been falsely recorded as offenders when they were in fact victims or witnesses, resulting in escalation of risk assessments (CO, 2020a, p. 17; 2020b, p. 42). The Ombudsman noted a ‘tendency for analysts to include inaccurate information’ (CO, 2020b, p. 17). SRAT risk ratings can be based on partial information; for example, without information about a person’s conduct in prison due to a lack of information sharing protocols (AHRC, 2019a, p. 22). It is also unclear whether Serco’s risk assessments consider other relevant information such as sentencing remarks, evidence of good behaviour in prison or other risk assessments undertaken at the bail or parole stage (AHRC, 2019a, p. 22). Further, risk assessments have been found to take account of irrelevant information. For example, Serco staff appear to have manually overridden Mr CE’s risk rating on the basis of his Vietnamese nationality because in the past Vietnamese detainees had attempted to escape from detention (AHRC, 2019b, p. 55). The flawed design and operation of risk assessments can amplify the level of risk attributed to detainees and result in the harmful and unreasonable use of force, and as the next part contends, such risk assessments are also relevant to an officer’s criminal culpability for harmful use of force.

The law of force in immigration detention

If an officer uses force in circumstances not justified by statutory or common law powers, that force might amount to a criminal offence, such as assault, punishable by a term of imprisonment. This is why some scholars contend police powers are better understood as limited exemptions from criminal and civil liability (Dixon, 1997). This part sets out the statutory and common law powers available to Serco security staff. It demonstrates how the criminal law makes assessment of risks, including that detainees will be violent to others, harm themselves, or escape, relevant to officers’ culpability.

The broadest statutory power to use force available to Serco staff in Australia authorised as officers under the Migration Act is arguably derived from the requirement to detain if the officer ‘knows or reasonably suspects that a person in the migration zone … is an unlawful non-citizen’ (Migration Act s 189(1)). Although the mandatory detention provision does not include explicit authorisation for an officer’s use of force, ‘detain’ means ‘take into’ or ‘keep, or cause to be kept in immigration detention’ and includes ‘taking such action and using such force as are reasonably necessary to do so’ (Migration Act s 5(1)). In ARJ17 v Minister for Immigration and Border Protection (2018) at [73] (‘ARJ17’) Rares J stated that the ‘power attached to the definition of “detain” is conditioned by what is, objectively, reasonably necessary, as an action or use of force, to detain or keep a person in detention’. The statutory power to detain ‘includes the power to … use such force as is reasonably necessary to keep places of immigration detention safe, ordered and secure’ (ARJ17 at [47]). For example, the power to detain entitles an officer to use force to remove a weapon from a detainee because the Migration Act makes it an offence for a detainee to possess a weapon, but would not justify taking action to remove a mobile phone from a detainee because such action ‘is not self-evidently reasonably necessary to keep’ a detainee in detention (ARJ17 at [71]-[72]).

Other provisions in the Migration Act similarly require an objective assessment of the circumstances for the lawful use of force. Officers ‘shall not use more force … than is reasonably necessary’ to conduct a search, strip search, or screening procedure, and may use ‘reasonable force’ to collect identifiers (see Table 3). Case law and literature on police powers is relevant to interpretation of statutory powers held by officers under the Migration Act. Courts interpreting Migration Act provisions which confer a power conditioned upon what is ‘reasonably necessary’ or upon ‘reasonable grounds’ for a state of mind have drawn on jurisprudence on police powers to aid interpretation (see for example ARJ17 at [73] and Ruddock v Taylor (2005) at [71], [84], [85] in relation to the power to detain). Courts interpreting statutory provision for police use of force, phrased similarly to the Migration Act provisions, have held that force must be proportionate to all the circumstances of the case (see for example Hamilton v State of New South Wales (No 13) [2016] at [171] ‘Hamilton’).

Table 3. Migration Act 1958 (Cth) provisions empowering the use of force.

• s 5 detain; s189 detention of unlawful non-citizens; s 192 detention of visa holders whose visas are liable to cancellation; s 253 detention of deportee

• s 198 removal of unlawful non-citizens from Australia

• s 252(8) pat searches of persons

• s 252AA(2) power to conduct a screening procedure

• s 252A strip search detainee in particular circumstances; s 252B(1)(k) use of force in conducting a strip search

• s 261AE(1)–(4) use of force in carrying out identification tests

Private security, such as Serco officers working in detention, also rely on the scope for individuals to lawfully use force. Criminal law statutes across Australia generally provide that a person will not be criminally culpable for using force in self-defence. While the law of self-defence differs between jurisdictions, generally a person acts in self-defence if they believe the conduct is necessary for one of the following purposes—to defend oneself or others, to protect property from destruction or damage, to prevent the unlawful imprisonment of themselves or others, to prevent criminal trespass to land—and the conduct is a reasonable response in the circumstances as they perceive them (see for example Crimes Act 1900 (NSW) s 418; Criminal Code Act 1995 (Cth) Criminal Code Schedule s 10.4(2)). Thus, a Serco officer’s subjective perception of the necessity of their conduct, and an objective assessment of the proportionality of the officer’s response to the situation the officer subjectively believed they faced is relevant (see R v Katarzynski [2002] at [10], [22]-[24]).

The ‘reasonableness’ of an officer’s use of force is thus crucial to determining whether an officer’s use of force is within the bounds of statutory authority or otherwise justified as self-defence. If it is not, then the use of force may be without lawful footing and, if the elements of the relevant offence are established, the officer may be found to be criminally liable. Yet statutory thresholds of ‘reasonably necessary’ and ‘reasonable grounds’ cannot be understood in isolation from law and policing in practice (Dixon, 1997, pp. 49–89). ‘Reasonable’ and ‘ordinary person’ standards have been found to reflect dominant understandings and viewpoints in a range of legal areas (Blore, 2018, p. 160; Douglas, 2006, p. 208). Reflecting on the barriers litigants face when challenging state actors in tort law for deaths in custody in the United Kingdom, Tuitt contends that the reasonable person test is susceptible to treating commonality as reasonableness and thus to being shaped by institutionally inscribed racial knowledge and norms (Tuitt, 2004, p. 40).

Such analyses suggest that legal thresholds of ‘reasonableness’ have a malleability which may favour treating force used in accordance with policy and procedural norms (such as SRAT and risk management practices) as reasonable, despite their flaws. On the one hand, evidence that police used accepted techniques on alleged offenders is not decisive to the evaluation of the reasonableness of officers’ conduct (Hamilton at [172]). On the other hand, an officer’s knowledge of the risk rating of a detainee at the time of forceful contact would seem to be a powerful factor. An officer’s knowledge at the time provides a factual basis to determine the reasonableness of an officer’s conduct even if later found to be inaccurate (see R v Rondo (2001)). Institutionalised risk assessment practices, however flawed, may thus limit the potential for the criminal law to operate as a mechanism for accountability for harm experienced by detainees at the hands of Serco officers.

Privatisation of immigration detention and the use of force

The privatised nature of immigration detention involves several characteristics that provide weak legal guidance and arguably incentivise Serco to use force inappropriately and based on flawed risk assessments. First, the Migration Act provisions for officers to use force (see Table 3) do not statutorily codify standards as is conventional, for example, in juvenile criminal legal detention (HREOC, 2004b, p. 878). The Human Rights and Equal Opportunity Commission explains:

Thus, the development of standards and systems for immigration detention has been left to the internal systems of the Department. In practice, this has taken place through commercial agreements with detention services providers, ad hoc arrangements with State authorities and Departmental guidelines. As detainees are not parties to such agreements, they have no enforceable rights to be treated in accordance with the standards and have no direct remedy for a failure to meet those contractual standards. (HREOC, 2004b, p. 878)

The lack of specific binding legal guidance on Serco’s use of force is typical of private security whose legal authority has been described as ‘determined more by a piecemeal array of legal privileges and assumptions’ due to neglect of legal arrangements for cooperative public-private activities (Sarre, 2008, p. 303).

Contractual regulation of immigration detention has varied over time, but almost two decades later the assessment above by HREOC (later renamed AHRC) remains true. Penovic (2014, p. 47) explains that the third phase of contractual regulation of private detention service providers, with Serco at the helm, adopts an ‘“outcomes-based” approach designed to give Serco operational flexibility in performing its contractual obligations’ involving ‘unprecedented power to monitor its own performance’.

As staff of a private company, Serco officers are not bound by the ethical rules governing the public service, however they are contractually bound to abide by the Department’s Detention Services Manual (DSM). The DSM (DHA, 2018) sets out that force must only be used as a last resort—officers must first consider conflict resolution through negotiation and de-escalation where practicable. It provides that force must not be used as punishment and should be proportionate to the situation and reasonable in the circumstances. Yet these policies and procedures do not have the status of law meaning that breach does not necessarily make the use of force a criminal assault. Note that policy guidance requires Serco to obtain prior approval for planned use of force from the ABF Detention Superintendent for that facility. Yet while additional requirements apply to planned use of force, the Ombudsman has expressed concern that, for example, requirements to obtain medical guidance prior to use of restraints have seemingly been circumvented by labelling of force as ‘unplanned’ despite non-urgent circumstances (CO, 2023, p. 21).

Second, the contractual framework arguably institutionalises incentives for high-risk ratings. In the past, contractual terms binding private immigration detention service providers imposed financial penalties for detainees’ escape which ‘created an incentive for a high-security environment in detention centres irrespective of the risk posed by individual detainees’ (JSC, 2008, pp. 66–67). Specific penalties for the escape of detainees have since been removed from contracts (JSC, 2008, pp. 66–67). However, contracts continue to utilise financial incentives and abatements to regulate contractual performance measures which in turn are mapped to key performance indicators (Commonwealth of Australia and Serco Australia, 2015, Schedule 4). Key performance indicators, such as ‘security’, suggest a potential incentive for high-risk ratings and the use of force or restraints in lieu of other means to avoid ‘incidents’ that may otherwise attract abatements for failing to ensure the security of detention.

Third, the longer controversy over the legal boundaries of Serco’s use of force in detention similarly suggests that Serco has an incentive to favour high-risk ratings in the design and operation of risk assessment to legitimise any use of force. As early as 2001, the Ombudsman queried the legal authority of officers to use force and restraints (CO, 2001, 7, pp. 28-29). In 2010 Serco obtained legal advice that its powers did not extend to restoring public order in detention (Hawke & Williams, 2011, p. 104). The Hawke and Williams (2011, p. 104) review likewise identified the need to clarify Serco’s role in providing good order in detention, and particularly the division of roles between Serco and the AFP or State and Territory police.

In 2015 the former Immigration Minister cited the Hawke and Williams report when he sought to introduce expansive powers for officers to use ‘reasonable force’ in detention and bar proceedings for force used in ‘good faith’ (Commonwealth, 2015; for discussion see AHRC, 2015; McAdam et al., 2015).⁴ The Bill proposed officers would have authority to use such reasonable force as the officer ‘believes is necessary’ to ‘protect the life, health or safety of any person (including the authorised officer)’ or ‘maintain the good order, peace or security’ of an immigration detention facility. The concept of ‘good order, peace or security’ was not defined, thus legal scholars noted ‘an officer might reasonably believe a peaceful protest by asylum seekers is a matter that “disturbs good order, peace or security” and use force to disperse those involved’ (McAdam et al., 2015, p. 3). While this Bill did not pass, there is some correlation between the proposed provisions and the SRAT. For example, an SRAT risk rating can be based on the number of times a detainee participated in demonstrations onsite, and thus may be utilised to justify an officer’s use of force.

In 2017 and 2020, again the Immigration Minister proposed to expand the police-like powers available to officers in ways that would reduce scrutiny by the criminal legal system (Boon-Kuo, 2022).⁵ Again, the proposed laws did not pass. It is my argument that the laws proposed in 2015, 2017 and 2020 sought to address what had become the ‘problem’ of the lack of clear legal boundaries for Serco’s use of force, and in the wake of their failure to pass, risk assessment and risk ratings have functioned to construct Serco’s actions as lawful.

The state is also invested in avoiding scrutiny of the policies governing Serco’s practice. The Department opposed the release of Serco’s Policy and Procedure Manual for three years (Taylor & Knaus, 2020). The Department explained that part of their reason for opposition was that it would aid several detainees regarded as having created ‘excessive complaints’ to external bodies in ‘their stated purpose of causing administrative burden to Serco and the department’ (ASRC and DHA (FOI) [2020] at [43]). The Department seems to have foregrounded corporate goals of administrative efficiency over grant of freedom of information requests, an approach that is consistent with theorisation of private detention centres as examples of the mutual constitution of law and power by public and private entities (Doty & Wheatley, 2013, p. 428).

Conclusion

Immigration detention has now been managed by private corporations for 25 years. There was a moment in 2007 when it looked like that would change. The Labor Party had run on the platform that it would return immigration detention to public hands, but soon after they were elected, the newly appointed Immigration Minister announced ‘a lack of alternative public service providers’ (cited in Penovic, 2014, p. 37). Penovic remarks ‘the Commonwealth appeared to have lost the necessary knowledge and expertise to manage immigration detention facilities’ (2014, p. 38). Today, part of the ‘knowledge and expertise’ in managing immigration detention involves Serco’s risk assessments of detainees. This article has argued that the flawed design and operation of risk assessment practices masks the seriousness of Serco’s harmful use of force and restraints on detainees. Force and restraints may be more readily used on those detainees assessed as ‘high-risk’, and that force comes to be perceived as within the boundaries of the criminal law. Risk assessment practices have resulted in instances found to be in breach of policy and human rights—such as the handcuffing of a compliant child, or the painful use of cuffs over an adult’s wrist wounds for over eight hours (AHRC, 2019b, p. 98, 50-51)—which have not led to criminal prosecution. The effect of risk assessments reflects the crucial character of immigration detention as argued by Peterie (2018, p. 306)—the capacity to ‘break’ people inside while maintaining the appearance of civility and respect for the law.

It is anticipated that changes will be made to the security risk assessment tool in response to the Griffith Institute of Criminology report, but at the time of writing in 2023 no amendments had yet been made (AHRC, 2023, p. 39). Should changes to the SRAT be made, it is likely that private corporations, presently Serco, will continue to be contractually responsible to conduct risk assessments for people in detention (AHRC, 2019a, p. 21). Thus, it important to question the legitimacy of the use of force in detention and practices which undermine accountability by fostering private actors’ power to shape the regulatory environment in ways favourable to their business. While Serco’s forceful practices are not always hidden, the mutual constitution of legal and institutional practices of risk assessment by Serco and the state mean that Serco officers’ use of force is constructed as lawful, and not criminalised, which ultimately sustain its violence.

Acknowledgements

This article has benefited greatly from discussion and guidance from the editors of this special issue, Penny Crofts and Elyse Methven, and the 2022 symposium they convened. I am also grateful for the generous engagement of Vicki Sentas and anonymous referees on an earlier draft of this article. Any flaws remain my own.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The Commonwealth Government has contracted the following corporations to provide ‘immigration detention services’—Australasian Correctional Management Pty Ltd (ACM) (1998–2003), Group 4 Falck Global Solutions Pty Ltd (G4S) which later changed its name to Global Solutions Limited (Australia) (GSL) (2003–2009), and Serco Australia Pty Ltd (2009–present) (Penovic, 2014, pp. 12–13).

2 The name of the Department of Immigration has changed multiple times over the period covered by this article, and is currently named the Department of Home Affairs. This article refers to all iterations as ‘the Department’.

3 The number of children is derived from monthly statistical reports on people in detention in the period 1 January 2016–31 December 2020. I selected the monthly report from each year which recorded the highest number of children held in detention.

4 Migration Amendment (Maintaining the Good Order of Immigration Detention Facilities) Bill 2015

5 Migration Amendment (Prohibiting Items in Immigration Detention Facilities) Bill 2017; Migration Amendment (Prohibiting Items in Immigration Detention Facilities) Bill 2020.