Abstract
The British state is currently taking forward deeply contentious legislation that would essentially end all legacy investigations and court cases relating to the conflict in the North of Ireland (1968–1998). Shaped by a long-term rightwing campaign to prevent any further investigation or prosecution of former British soldiers, and a wider culture of denial of the role of state collusion in the conflict, the legacy proposals are ostensibly defended on the grounds that current mechanisms do not work for victims’ families. This article seeks to both challenge that narrative and to build on earlier analyses of collusion (Race & Class 57, no. 2; 58, no. 3) to demonstrate how recently published reports of official investigations into collusion between state agents and loyalist paramilitaries have provided important information for victims’ families and insights into the patterns of collusion. Such patterns can be identified in terms of state actions and omissions taking place before, during and after lethal loyalist attacks. They include providing weapons and targeting intelligence while failing to provide warnings to those being targeted; the direct involvement of serving and former members of the security forces in loyalist killings; blocking investigations, destroying records and employing (and protecting) state agents and informers involved in mass murder.
Introduction: dealing with the past
In May 2022 the British government introduced the controversial Northern Ireland Troubles (Legacy and Reconciliation) Bill dealing with the legacy of the conflict in the North of Ireland which cost some 3,600 lives (and 30,000 injured) between 1968 and 1998. 1 Centred on the creation of a new, judge-led Independent Commission for Reconciliation and Information Recovery Commission 2 that will undertake ‘reviews’ (rather than investigations) of the 1,500+ unsolved conflict deaths, the Bill also proposes to end all other investigative and legal avenues for truth and justice for victims and relatives. 3 In an unprecedented step, and challenge to the foundations of due process and the rule of law, the legislation will stop virtually all further conflict-related criminal and civil cases, inquests and any further investigations undertaken by independent watchdog bodies such as the Office of the Police Ombudsman for Northern Ireland (PONI), created in the wake of the Good Friday Agreement (GFA). Equally contentious, the new Commission will also have the power to grant what amounts to an amnesty for those whom it deems to have ‘told the truth’ about any conflict-related killing. Tellingly, however, amnesties will not apply to anyone who has been convicted for Troubles-related offences – the overwhelming majority of whom are the very former members of republican and loyalist armed groups most likely to have information about many unsolved deaths. Against the background of a long-term campaign waged by the British military and political establishment and the rightwing press, amnesties are essentially designed not to garner truth for families, but to prevent any further investigation and prosecution of former British soldiers for state killings.
The British government plan has been met with vociferous and wholesale opposition. Victims’ organisations have universally opposed the legacy bill as have (for various reasons) all the major political parties in the North, the Irish government, the US administration and human rights agencies. The Northern Ireland Human Rights Commissioner declared the bill as ‘fatally flawed’, incompatible with the European Convention on Human Rights and the 1998 Human Rights Act. 4 The Council of Europe Commissioner for Human Rights has stated that the ‘unconditional amnesty’ contained in the proposals would lead to ‘impunity’ and see Britain likely to be in breach of its ‘international [legal] obligations’. 5
Flying in the face of such opposition, a central platform of the government’s case has been that it is acting in the interests of victims because the ‘current system’ and ‘divisive legal processes . . . [are] not working for anyone’. 6 Certainly victims’ families have faced an uphill struggle to obtain truth and justice via courts and investigations. That has often been the result, however, of the obstructionism and obstacles put in the way by the government, Ministry of Defence (MoD) and other state agencies, including the police, intelligence services and the British Army. This has been particularly so for the relatives of those killed by state forces or where there have been longstanding allegations of collusion between state agencies and (particularly loyalist) non-state violent organisations. Despite such hurdles, and the limitations of these processes, important information has been revealed in terms of the extent and nature of collusion in many conflict killings. That, rather than an inherent belief in British justice, is one reason why the victims of state violence have been so vocal in their condemnation of the closing down of these means to seek truth.
The aim of this article is in part therefore to challenge the state contention that legacy investigations do ‘not work for anyone’ by showing how – in terms one of the most disputed aspects of legacy and the Troubles – recent investigations have increasingly shed a light on allegations of state collusion with loyalist paramilitaries. Primarily, however, the article is designed to illuminate the patterns of collusion that such investigations have helped to reveal. The main focus will be several recent reports of investigations undertaken by the Police Ombudsman, Marie Anderson, into allegations of state collusion in dozens of loyalist killings of nationalists and republicans in the late 1980s and early 1990s in various parts of the North. 7 What these reports do is to fill in more pieces of the jigsaw that help to reveal the full picture of what collusion involved. It is the patterns of collusion that matter. Evidencing collusion in this way can also substantiate the character of state wrongdoing and provide a detailed paradigm that may find echoes elsewhere.
Such patterns might be identified in a timeline evident before, during and after attacks. This article will be organised in that vein. It first briefly contextualises what we might understand collusion to be, the record of investigations into collusion and the scale of loyalist violence that makes the late 1980s and early 1990s such an important period to examine. Parts two and three look at the evidence of pre-attack collusion that emerges from recent investigations. First, in terms of supplying loyalist groups with weapons and then the failure to provide warnings to targeted members of the nationalist community. Evidence of serving and former members of the security forces who were directly involved in killings carried out by loyalists is explored in part four. Part five examines the ways in which investigations into loyalist killings were undermined both at the time and after, not least (as explored in part six) by the destruction of records – something which also limited the ability of the recent Ombudsman inquiries to get to the full truth and ensure accountability. Finally, the article looks at evidence of collusion these reports reveal in terms of the ways that agents and informers were handled by state agencies – an issue that lies at the heart of collusion as institutionalised and systemic state practice. The piece concludes by considering how such evidence of collusion challenges an ongoing culture of denial.
Collusion and loyalist violence
Collusion here refers to the involvement of state agents (members of the police, army, prison and intelligence services) or state officials (government ministers, legal officers, civil servants), directly or indirectly, through commission, omission, collaboration or connivance, with armed non-state groups or agents, in wrongful acts usually (although not exclusively) involving or related to non-state political violence and extra-judicial killing. 8 There have now been numerous official reports, published over decades, confirming longstanding allegations of collusion between state agents and loyalists taking place throughout the conflict in the North of Ireland, and particularly from the late 1980s to the early 2000s. They include the 2003 Inquiry undertaken by former Metropolitan Police Chief Sir John Stevens, the reports published by retired Canadian judge Peter Cory in 2004 and the 2012 de Silva review into the killing of human rights lawyer Pat Finucane. 9
Geographically, collusion with loyalists has now been officially found to have occurred in areas right across the North, throughout different periods of the conflict and in some of the worst atrocities of the ‘Troubles’. The notorious ‘Glennane Gang’, which carried out a vicious sectarian killing campaign in Armagh and Tyrone in the 1970s, included numerous serving and former members of the British Army and Royal Ulster Constabulary (RUC) in its ranks. 10 Its actions − including the 1974 Dublin-Monaghan bombings (which saw the largest loss of life on a single day throughout the conflict) − remain the subject of an ongoing investigation under the auspices of Operation Kenova. 11 Families continue to pursue various routes to justice over longstanding allegations of collusion with loyalists in dozens of killings in Tyrone and Mid-Ulster through the late 1980s and early 1990s. Collusion was involved in the 1994 Loughinisland massacre in Co Down, when six Catholic men were shot dead and five seriously wounded in O’Toole’s bar by members of the UVF. 12 There was RUC collusion with loyalists, ‘up to and including murder’, in North Belfast through to the early 2000s. 13 It is telling that parallel patterns of collusion emerge across time and space.
Successive police ombudsman reports have found evidence of widespread and systemic collusion in dozens of killings. Two of the most recent – both published in 2022 – form the main focus here. 14 They are the result of investigations into collusion between members of the RUC and loyalist paramilitaries in the north west area (centred on South Derry) and in South Belfast. These followed extensive and prolonged investigations into a total of thirty-one murders and three attempted murders – linked thematically by collusion allegations – that occurred in these areas between 1989 and 1998. All but two of these attacks took place in an intense period of loyalist violence between 1989 and 1994. Nineteen killings were carried out by the North West Ulster Defence Association (UDA) under its cover name of the Ulster Freedom Fighters (UFF). 15 These included the mass shooting of four Catholic workmen in Castlerock on 25 March 1993 and the Greysteel massacre, when a total of eight people were killed in a UDA/UFF gun attack on the Rising Sun bar on 30 October 1993. 16 Twelve people were killed by the South Belfast UDA/UFF, including the five victims (one a 15-year-old boy) of the gun attack on Sean Graham Bookmakers shop on the lower Ormeau Road on 5 February 1992. A third recent report discussed examines collusion in the 1993 killing of Catholic teenager Damien Walsh in West Belfast. 17
The early 1990s saw a massive escalation of loyalist violence. This occurred against a backdrop of an emerging ‘shoot-to-kill’ policy that saw a huge increase in the number of active republicans killed by state forces from the late 1980s onwards − particularly through ‘set-piece ambushes’ or ‘selective assassinations’ carried out by ‘specialist’ British Army units (such as the SAS) and counterinsurgency units within the RUC. 18 The rise in loyalist killings in this period was in stark contrast to the years immediately preceding. For example, despite a backdrop of widespread unionist/loyalist mass opposition to the signing of the Anglo-Irish Agreement, 1985 saw the lowest number of people killed by loyalists since 1969 (four). 19 In the next five years loyalists claimed a total of ninety lives. However, in the four that followed (1991–1994), that figure almost doubled, to 164. In 1991 alone loyalists killed forty people – twice as many as the year before and a scale of lethal violence not seen since the mid-1970s. By 1993 loyalists were responsible for more killings than anyone else, including the IRA.
The rapid increase of loyalist attacks is mirrored in South Belfast and in the north west. For example, in the three and a half years up to 1990, the UDA had killed twenty-one people throughout Belfast, and four in South Belfast. Between late 1990 and early 1994 the UDA killed fifty-six people in the city as a whole, including twenty victims in South Belfast alone. 20 State forces were aware from 1989 onwards, that loyalists were intent on ramping up their violence and that the targeting of republicans was central to their plans. In the north west this fed into concerted efforts to increase recruitment of informers within loyalist ranks. Infiltration of the north west UDA ‘from top to bottom’ would not, however, prevent a marked increase in loyalist violence here too. 21
Providing weapons
The capacity of loyalists to significantly increase their violent campaign in the early 1990s depended on having access to a new arsenal of weapons. These came from several (often police or military) sources, but a crucial part was played by a large cache imported from South Africa in December 1987. 22 There were longstanding allegations of British military intelligence and RUC Special Branch collusion involved in this shipment. They centred on the role of Brian Nelson, a Belfast-native and former member of the British Army who, by the late 1980s, was Chief Intelligence Officer of the UDA/UFF. At one and the same time he was also an agent working on behalf of the Force Research Unit (FRU) – the key British Army intelligence unit operating in the North.
Collusion allegations in the South African arms shipment were examined as part of the 2012 de Silva review into state involvement in the loyalist murder of civil rights lawyer Pat Finucane in North Belfast in 1989. While de Silva found copious evidence of collusion in many aspects of Pat Finucane’s death, he concluded that neither Nelson nor the FRU had any prior knowledge or involvement in the arms shipment. Arguing, indeed, that the RUC capture of parts of this huge weapons cache was ultimate proof loyalists were not ‘simply state-sponsored forces’. 23 However, the 2016 Loughinisland report found otherwise. 24 Contrary to de Silva’s findings, then Police Ombudsman Michael Maguire found considerable evidence that RUC Special Branch and British Military Intelligence knew a great deal about the South African arms shipment and singularly failed to stop a significant number of these weapons falling into loyalist hands and their use in over 80 murders. 25 These guns were a primary means through which loyalist killing accelerated through the next few years. This was as true in South Derry and South Belfast as it was elsewhere. For example, South Belfast, the semi-automatic rifle used in the Ormeau Road massacre had come from the South African cache of guns. 26
In the north west, if anything, the picture was even starker and also reveals the significant part played by another loyalist organisation – Ulster Resistance (UR). While both the UDA/UFF and UVF were deeply involved, UR was central to the 1987 arms shipment. Founded in 1986, by the leadership of the DUP, UR was conceived as a mass-based ‘clean-living paramilitary group’, through which ‘loyalist politicians and business elements sought to harness the political potential of paramilitary muscle’ in opposition to the Anglo-Irish Agreement. 27 Senior DUP figures occupied leadership positions. Former and serving security force members were included in its ranks. 28 So too were leading loyalists, including John McMichael, head of the UDA and ‘mentor’ to British agent Brian Nelson. McMichael had been central in developing arms links in South Africa. It has also been suggested that ‘for many years’ he had been a ‘useful conduit for acting on British Intelligence targeting information and may have been one among many informers and agents operating at senior levels in the UDA/UFF’. 29
It was Ulster Resistance which was the prime mover in providing assault rifles from the arms shipment to the North West UDA/UFF. 30 The Ombudsman found evidence that Ulster Resistance demanded payment ‘at current prices’ for weapons it supplied to the UDA/UFF and UVF to replace those the latter had lost in police finds. The North West UDA clearly found the means to do so, taking delivery of a stock of pistols and up to five assault rifles from the consignment in the first half of 1988. 31 They were soon being put to deadly use. One of the South African AZ58 rifles was used to shoot dead Gerard Casey as he lay in bed beside his wife and 3-month-old child at their home in Rasharkin, South Derry in April 1989. 32 Another was used to kill Patrick Shanaghan in West Tyrone in August 1991. 33 A Browning pistol likely to have been part of the 1987 shipment was used to kill Sinn Fein councillor Eddie Fullerton and Danny Cassidy. 34 The latter had been subject to a very public campaign of harassment by a specialist RUC unit before he was killed near his home in Kilrea, South Derry in April 1992.
The same South African semi-automatic rifle used to kill Gerard Casey was the main weapon employed to carry out the Greysteel massacre on 30 October 1993. 35 One of the most notorious, indiscriminate mass killings of the conflict, the Halloween night ‘trick or treat’ attack (so-called because that is what the gunman shouted as he opened fire) on the Rising Sun Bar in Greysteel left eight people dead and ten others seriously wounded. Eight months later, in parallel vein, but on the opposite side of the northern state, another South African shipment VZ58 would be used to slaughter six men as they watched Ireland play a world cup football match at the Heights bar in the small townland of Loughinisland. 36 Collusion and the South African weapons provided loyalists with means to kill right across the North of Ireland, and to both target republicans on the one hand, and terrorise the wider nationalist community on the other.
Allowing the importation of a large cache of guns was not the only way in which loyalists accessed weapons through state collusion. Others included allegedly facilitating ‘thefts’ from state armouries, and – extraordinarily – in returning seized weapons to loyalist killing squads. It is worth examining each in turn. The ‘theft’ of police and military weapons formed an important source of arms for loyalists. For example, a revolver used in the killing of Theresa Clinton in South Belfast had reportedly been stolen from an RUC officer in December 1991 and already used in three other killings, and four attempted murders. 37 A Browning pistol used in the Ormeau Road massacre had been one of four weapons supposedly stolen from Malone UDR barracks in January 1989. 38 In this case, the guns had actually simply been issued to two men dressed in British Army uniforms and had then ‘disappeared’. UDA/UFF informer Ken Barrett, pivotal in the killing of Pat Finucane, carried out the ‘theft’. 39 Investigated by the Stevens Inquiry, the de Silva Report revealed that Barrett later claimed the ‘theft’ had been organised by his RUC Special Branch handler. 40
This directly parallels the history of another Browning pistol used in several loyalist murders, including that of human rights lawyer Pat Finucane in his North Belfast home in February 1989. 41 This weapon was one of a number (including machine guns) that had been sold by a UDR Colour Sergeant to British agent Ken Barrett in August 1987. 42 Barrett later said an RUC officer phoned the UFF to tell them when the weapons were ‘on their way’. The UDR soldier was himself a former member of the UDA who had two personal weapons previously ‘stolen’ by the UDA on separate occasions in the 1970s. His military career did not suffer as a result. 43 By November 1988 this Browning pistol was in the hands of Special Branch informer William Stobie − information he happily shared with his handlers who did nothing to stop him handing weapons over to other UDA members involved in killings. Indeed, Desmond de Silva found that the RUC only stepped in to covertly disrupt Stobie distributing guns when warned of a possible UDA attack against the police themselves. 44
‘Stolen’ state weapons were prominent in the killings carried out by the North West UDA/UFF too. A revolver claimed to have been stolen from a former RUC officer was used in the lethal attacks on both Sinn Fein councillor Eddie Fullerton in Donegal and republican former prisoner Thomas Donaghy in Kilrea in south Derry in the summer of 1991. 45 A handgun used to kill Sinn Fein councillor Bernard O’Hagan in September 1991 and republican former prisoner Malachy Carey in December 1992 had been reportedly stolen from a UDR man as long ago as 1975. 46 Another pistol supposedly stolen from the boot of a UDR soldier in March 1992 was one of the guns used to kill the four men in the Castlerock massacre a year later. 47
An even less well-known aspect of weapons collusion, is that guns seized by the authorities were, in some circumstances, handed back to loyalists and then used in killing Catholics. The ‘stolen’ Browning pistol used in the Ormeau Road massacre was also one of a number of weapons that had been returned to loyalists after having been ‘seized’ by the state authorities. Barrett had given the gun to William Stobie, who, as well as being an RUC informer, was also quartermaster for the UDA in West Belfast. 48 Along with several other weapons, Stobie gave the Browning pistol to his handler in November 1989. This was the same police officer Barrett accused of organising the theft of the weapon from the UDR barracks in the first place. Days later, this RUC man gave the guns back to Stobie. 49 A follow-up search intended to again recover the Browning pistol failed to do so. It was used to kill 22-year-old Aidan Wallace in December 1991 before being used in the Ormeau Road attack less than two months later. 50 Aidan Wallace died when the UDA/UFF fired indiscriminately into a crowded bar on a Sunday lunchtime. Three others, including an 8-year-old boy, were wounded.
It appears a ‘specialised policing unit’ within RUC Special Branch was involved in handing back weapons to paramilitary groups throughout the 1980s. 51 Before being returned, such weapons were supposed to have been ‘deactivated’. In theory, formal authorisation for such actions was required from senior RUC officers. In practice, things apparently worked on a much looser, more ‘ad hoc’ (and so less readily accountable) basis. ‘Decision-making processes were verbal’, the Head of RUC Special Branch told the Stevens Inquiry, ‘and not documented’. 52 For the Police Ombudsman, the ‘deliberate failure’ to document how and when weapons were given back to loyalists – making ‘scrutiny, then and now, impossible’ – showed a ‘desire to avoid accountability’ and represented a textbook case of collusion as the ‘wilful failure to keep records’. 53 Giving weapons back to agents like William Stobie also ‘demonstrated a disregard for the safety of members of the public by the police’.
While in police hands the Browning pistol was ‘deactivated’. Some of the other returned weapons were not. In any case, it is now clear that RUC Special Branch were also aware that the UDA in Belfast had the knowledge and means to ‘reactivate’ such weapons. 54 That is what happened to this gun. It was not the only ‘reactivated’ weapon used by the UDA in South Belfast. A revolver used to kill Martin Moran as he delivered for a Chinese takeaway on 23 October 1993 (less than 12 hours after the Shankill Road bombing) was a ‘reactivated’ weapon. A sterling sub-machine gun used to kill Theresa Clinton had also previously been officially ‘de-activated’. The Ombudsman could not account for how this weapon ended up, ‘reactivated’ and in the hands of the UDA. 55 RUC Special Branch did not inform those investigating this murder (and that of Martin Moran) that the UDA had the capacity to reactivate such weapons, or knowledge of who was involved. 56
The Browning pistol used in the Ormeau Road massacre was recovered shortly after by the police. The team investigating the killings was told it was a stolen weapon. However, Special Branch did not tell them about the circumstances of the theft. Even more important, it did not tell the investigators that the gun had been in their possession before being handed back to Stobie. Nor did it tell them about the movement of the weapon after it had been returned to the UDA. This potentially vital ‘history’ of the weapon was, concludes the Ombudsman, ‘deliberately concealed’ to protect Special Branch informers. 57 The gun was subsequently given back to the British Army – losing any further forensic opportunities as a result. Precisely the same thing happened to the Browning pistol used to kill Pat Finucane. 58
Accessing intelligence, targeting republicans
Desmond de Silva found that 85 per cent of all the intelligence held by loyalists by the late 1980s originated from state files. 59 Having victims’ personal details allowed loyalists to target people for assassination in a way not available to them before. FRU agent Brian Nelson was crucial here. Nelson himself provided state intelligence to the South Belfast UDA/UFF. For example, in September 1988 (with the full knowledge of his handlers) Nelson copied the personal intelligence files of ‘50 republican activists’ in South Belfast and passed them on to the UDA in the area. 60 However, as well as direct involvement in passing on state files from his army handlers used in targeting loyalist victims, Nelson was also central in developing the system of targeting intelligence developed by loyalists throughout the North. As de Silva concluded, Brian Nelson’s ‘express purpose’ was that other UDA areas should adopt the intelligence and targeting systems he developed (with the help of his handlers) in Belfast. 61
This is evident in the increased UDA/UFF access to state files in the north west that fed directly into their accelerated killing campaign. The North West UDA/UFF ‘significantly enhanced’ their intelligence capacity in the late 1980s and early 1990s. 62 Building up profiles of nationalists and republicans – particularly members of Sinn Fein and the IRA – ensured they had a ‘comprehensive and current intelligence picture’ when planning UDA attacks. 63 State files provided the major source for this targeting information. For example, in November 1989 a major cache of intelligence files ‘collected . . . over several years’ was found at the home of a known loyalist. 64 Gathered for ‘the sole purpose of facilitating UDA/UFF targeting for attacks’ most of the documents originated ‘from military regiments [including] army intelligence reports’. 65 It was one of a number of such finds. 66 Throughout South Derry and mid-Ulster loyalists had little previous success in targeting republican activists. That changed after 1989. This included killing Sinn Fein representatives and workers. During three decades of conflict, twenty-five Sinn Fein elected officials and members were killed throughout the North. 67 Of those 56 per cent (fourteen) were killed between 1989 and 1993 − and half of those deaths occurred in mid-Ulster.
UDR members were crucial in this supply of intelligence to loyalists. Perhaps the most striking example concerns ‘Person V’. A senior member of the UDR in the north west, ‘Person V’ attended meetings with the RUC ‘where sensitive information was discussed’. 68 This included ‘high level, strategic intelligence’ given to the UDR by RUC Special Branch. 69 Such intelligence was used in the killing of Gerard Casey in April 1989. 70 The murder of Gerard Casey was a pivotal moment. He was the first person killed by the North West UDA/UFF for several years. 71 Gerard Casey was a member of the IRA but that was far from public knowledge till after his death. 72 Along with the UVF killing of veteran republican and Sinn Fein councillor John Davey a month earlier, the death of Gerard Casey marked the onset of the new targeted campaign in South Derry in which state intelligence played a critical role. 73 RUC Special Branch concluded ‘Person V’ had been involved in passing information to the UDA/UFF, including that used to target and kill Gerard Casey. 74 Indeed, ‘Person V’ was subsequently dismissed from the UDR in late 1989. However, the RUC did not investigate his links to the UDA/UFF. The team investigating the killing of Gerard Casey were never told of ‘V’s suspected involvement and he was never investigated or arrested as a result. 75
Threat to life: a failure to provide warnings
By the late 1980s an ‘RUC Force Order’ was in place that required senior police officers to ensure a threat warning was issued to anyone whose life they had reason to believe was in ‘imminent’ danger. 76 Even a more general threat to life was supposed to lead to an assessment of the risks a potential victim faced and the possible issuing of a threat warning. However, what emerges from several reports now undertaken in various parts of the North confirms long held allegations that in many cases – and primarily for nationalist victims of loyalist violence − this simply did not happen. There are numerous instances where warnings were not given by the RUC when they should have been. Apart from anything else, this deprived people later killed by loyalists of the ability to take precautions to try to protect their own lives. There were many occasions, too, when the Ombudsman could find ‘no rationale’ for why these victims were not told of the threat they faced. 77
The de Silva report had already catalogued that, time and time again, intelligence received by RUC Special Branch of loyalist threats to the lives of nationalists and republicans was wilfully ignored, hidden and not acted upon. 78 Textbook evidence of collusion. In Belfast, information was not recorded in an RUC ‘threat book’ and warnings were not given to those being targeted. To give some idea of scale, threats from loyalists made up less than 5 per cent of the total recorded between 1987 and 1989. In the same period, loyalists were responsible for 45 per cent of paramilitary killings in Belfast. 79 Throughout the North, loyalists were responsible for 25 per cent of all killings between 1988 and 1989. Yet loyalist threats made up only 4 per cent of all those recorded. 80 Tellingly too, while loyalists overwhelmingly targeted and killed civilian members of the nationalist community, over 70 per cent of the loyalist threats that were recorded were of threats to members of the security forces. In other words, in Belfast and elsewhere, the RUC almost entirely ‘failed’ to record loyalists’ threats to nationalist civilians, never mind the republicans they were increasingly intent on targeting. Those specifically considered to be a ‘thorn in the side’ of the security forces, Desmond de Silva found, ‘were not provided with protection during this period of the troubles’. 81 It was an attitude shared by RUC Special Branch, Military intelligence and MI5 alike. As a result, ‘loyalist agents’ were ‘permitted to participate in criminal conspiracies’ and their intelligence not acted upon, ‘because the conspiracies related’ to those regarded as the state’s enemies. 82 It was a worldview and practice that meant, in the lead up to Pat Finucane’s killing, the RUC did nothing with the intelligence they received from William Stobie ‘regarding weapons, targeting or the UDA members involved in the West Belfast hit team’ who shot him dead. 83
These patterns were readily reflected in South Belfast, South Derry and the north west. In South Belfast, for example, the Ombudsman found that, in a number of killings, the police were aware of threats to the victims and knowingly failed to act. 84 In some cases, this was because RUC Special Branch did not pass threat intelligence on to others. 85 In October 1990, loyalists attempted to shoot dead Sam Caskey as he walked towards his parents’ home off the lower Ormeau Road. In 1988 and 1989 the RUC had received intelligence, on several occasions, that Sam Caskey was being targeted by loyalists. No ‘risk assessment’ was done and no threat warnings were given. 86 Sam Caskey also claimed that he had previously been threatened by RUC officers, both while held in Castlereagh interrogation centre and on the streets. 87 The RUC also received threat intelligence that loyalists were planning to attack Jim Clinton. Theresa Clinton, his wife and mother of two, was killed when loyalists shot into the family home, also just off the Lower Ormeau Road, on 14 April 1994. 88 Jim Clinton had received a threat warning from the RUC in 1989. However, as the RUC had intelligence on the ‘real and imminent threat’ to him and his family through early 1994, no risk assessment was done by the police, nor did they provide a warning. The Ombudsman concluded this constituted collusion. 89
As noted, in the north west the RUC uncovered several loyalist caches of state intelligence files used for targeting victims. While threat warnings for those being targeted should have followed, in many cases they did not. 90 This was noticeably so where republicans were concerned. Eddie Fullerton was a Sinn Fein local councillor in Donegal. He was shot dead in his home by loyalists on 25 May 1991. The only loyalist killing during a ten-week ceasefire, such a cross-border loyalist shooting was also highly unusual. 91 Information on Eddie Fullerton had been found among the cache of loyalist-held intelligence files discovered by the RUC in November 1989. Across his photo was written ‘dead as doornails’. 92 No threat warning was given by the RUC or An Garda Síochána (the southern Irish police force).
Collusion in terms of a failure to provide threat warnings is even more starkly illustrated in other cases in South Derry. Tommy Donaghy was a republican former prisoner shot dead by loyalists in Kilrea, Co. Derry in August 1991. 93 Exactly one month later, Sinn Fein councillor Bernard O’Hagan was also shot dead by loyalists as he arrived for work as a lecturer at Magherafelt FE College. 94 In the previous February, files on some 250 people had been found in a house in Portrush. The files included details on both Bernard O’Hagan and Tommy Donaghy’s family, 95 and most were of ‘military origin’. They were found in the home of ‘Person J’, a former British soldier who then joined and was still serving in the UDR in the north-west until early 1991. 96
However, an RUC Assistant Chief Constable decided that none of those whose names appeared in the recovered intelligence files should be given a threat warning because the documents had not ‘fallen into the hands of loyalist paramilitaries’, 97 despite the fact that the recent resignation of ‘Person J’ from the UDR ‘coincided with increased intelligence linking him to North West UDA/UFF activities’. 98 ‘J’ had been arrested following the files find, but released on bail and was not tried (and found guilty) for possession of the files ‘likely to be of use to terrorists’ until September. In the time between, Tommy Donaghy and Bernard O’Hagan were killed. Neither had been given threat warnings by the RUC. Even after these killings, no-one else listed in files the courts deemed ‘likely to be of use to terrorists’ were given a threat warning.
On 12 January 1992 the RUC found a gun, ammunition and explosives at the home of a former UDR soldier and suspected UVF member in Ballymoney. The names of ‘suspected’ IRA members were in a UDR notebook also recovered. Among them was the name of Danny Cassidy. No threat warning was given to Danny Cassidy by the RUC. 99 Nor, it appears, was a warning given to Patrick McErlain whose name also appeared in the notebook. A member of a well-known republican family, the UDA/UFF shot Patrick McErlain as he drove near his home in Dunloy, Co. Antrim on the morning of 28 August 1992. 100 The attack closely paralleled the killing of Patrick Shanaghan (see below). The same South African VZ58 rifle was used in both shootings. 101 Seriously wounded, Patrick McErlain nevertheless survived. It was not the first time the RUC had found loyalist-held files with details of Patrick McErlain. His name had already appeared, alongside that of Eddie Fullerton and others, on the UVF death list found by the RUC in November 1989. This cache of documents included targeting information on a total of thirty-one people, taken from military files and ‘army intelligence reports’. 102 It included the name of Gerard Casey, who had by then already been killed, ‘indicating that the individuals on the list were being targeted’. 103 Special Branch adjudged there was no ‘imminent threat’ to the others named. A loyalist cache of files had also been found in Belfast in November 1991. Patrick McErlain’s details appeared here too. No threat warnings were given to him after any of these finds. 104 Families’ ‘concerns about collusive activity’ in terms of a failure to warn a number of victims of ‘threats to their life’, the Ombudsman concluded, were ‘legitimate and justified’. 105
Indeed, even when threat warnings were provided, the attitude and response of the RUC could still prove deeply troubling. The killing of Patrick Shanaghan is a case in point. A member of Sinn Fein, Patrick Shanaghan was shot dead by loyalists as he drove to work on the morning of 12 August 1991. 106 Loyalists had tried to kill him once before, in February 1989. 107 But loyalists were not the only source of concern. Prior to his death, Patrick Shanaghan had been subjected to an extended campaign of harassment from members of the security forces. Between 1985 and 1991 he had been arrested and questioned by the RUC on ten separate occasions and held for a total of forty-two days. His home had been searched sixteen times. 108 Nothing was ever found and he had never been charged with any offence. Patrick Shanaghan had lodged several formal complaints about abusive treatment and death threats he had received at the hands of RUC interrogators. His statements record his having been physically abused, kicked, beaten and forced to stand in stress positions for ‘hours on end’. 109 One interrogator told him ‘an SAS type person [is] looking at you through the peep-hole this morning’, another that ‘loyalists in Castlederg know you now and they will get you’. 110 Witnesses testified to a catalogue of abuse and threats levelled against him when in custody. 111 He had, though, twice been given official warnings of threats from loyalists. The first was in December 1990 after intelligence files containing his details had ‘gone missing’ from the back of a British Army patrol vehicle. 112 Then, in April 1991 the RUC told Shanaghan ‘he was being targeted by loyalists’. 113 However, it emerged at the inquest that the RUC officer who had delivered this warning, far from advising him on steps for his safety, stopped and searched Patrick Shanaghan in his van only half an hour later. 114 As British-Irish Rights Watch noted shortly after, such warnings often ‘look more like threats than any attempt to protect the victim’. 115 He was shot dead four months later.
A 2021 report into the killing of Damien Walsh in West Belfast casts light on another dimension of how the RUC approached clear and evident threats to the wider nationalist community. 116 Damien Walsh was shot dead by members of the notoriously sectarian ‘C’ Company of the UDA at the shopping centre where he worked on the evening of 25 March 1993 – the very same day as the Castlerock massacre. The gun used was part of the South African shipment. 117 An Ombudsman investigation discovered that three days before the shooting the RUC suspended surveillance of the West Belfast UDA. Between then and the killing of Damien Walsh this loyalist gang had already shot dead a Catholic man, Peter Gallagher, and attempted to murder two other people with grenades − including a Sinn Fein councillor in Ardoyne. 118 Despite this, and receiving ‘multiple pieces of intelligence’ that ‘C’ Company was planning further sectarian attacks (and that its leader had received two handguns), the surveillance was not put back in place until five days after Damien Walsh was killed. It appears no ‘risk assessment’ was done, even after attacks were clearly happening. The RUC defence was that it had transferred resources for surveillance of the IRA. The ‘deliberate decision’ not to reinstate surveillance of the UDA in this period ‘allowed [loyalists] greater scope to mount terrorist attacks on the nationalist community’. 119
State forces involved in loyalist killings
There is also considerable evidence of members of the state security forces joining loyalist paramilitary groups and directly taking part in killings. For example, in one of the most notorious examples of collusion throughout the conflict, numerous serving and former members of the Ulster Defence Regiment (UDR) and Royal Ulster Constabulary (RUC) were part of the ‘Glennane Gang’. 120 This unit of the UVF was responsible for up to 120 murders in Armagh and Tyrone in the 1970s, including the 1974 Dublin-Monaghan bombings and mass killings in Catholic bars and villages. 121 More indirectly, there have also been allegations of elements within the security forces providing loyalists with clear routes in and out of areas to undertake attacks.
A network of serving and former members of the state security forces involved directly in loyalist paramilitary groups is evident throughout different parts of the North in the late 1980s and early 1990s. The recent reports on the actions of the UDA/UFF in the north west and South Belfast reflect this too. They include some familiar figures. As well as being the UDA Quartermaster for West Belfast, and an RUC informer, William Stobie was a former British soldier. 122 So too, of course, was the Head of UDA intelligence and FRU agent Brian Nelson. Others are less well-known.
Here, the case of ‘Person J’ is worth looking at in some depth. As noted earlier, ‘Person J’ was a former British soldier who was serving in the UDR in the north west until shortly before his arrest following an RUC loyalist files cache find at his home in February 1991. 123 However, this was not the first time he had been linked to loyalist paramilitaries. On 3 April 1989, the night before Gerard Casey was shot dead, ‘J’ was stopped by the RUC ‘acting suspiciously’ near an abandoned house where the getaway car used in the attack on Gerard Casey was later found. 124 He was accompanied by ‘Person K’, who was later suspected of being involved not only in the killing of Gerard Casey but in numerous others, including both the Castlerock and Greysteel massacres. 125 The police questioned both men in the days after the shooting. However, as a serving UDR soldier, ‘J’ was interviewed as a witness, rather than a suspect, 126 which ‘may have impeded the RUC investigation . . . [and] led to the loss of important evidential opportunities’. 127 Another UDR soldier provided his alibi. The February find of military files at his home led to ‘J’s conviction in September 1991. Unbelievably, he had claimed to have inadvertently ‘found’ the files (on over 250 people) in a plastic bag while on UDR patrol. 128 He received a one-year sentence.
Out on bail, several UDA/UFF killings took place in South Derry and the north west between ‘J’s arrest and conviction. They included the killing of Sinn Fein councillor Eddie Fullerton in his home that May. ‘J’ was suspected of being involved in this attack. 129 Former republican prisoner and Sinn Fein worker Tommy Donaghy was shot dead by the UDA/UFF as he arrived for work in Kilrea, Co. Derry early in the morning of 16 August. Echoing the killing of Gerard Casey, ‘J’ was stopped by the police a couple of days prior to the attack on Tommy Donaghy, at a similar time and near where he would later be killed. 130 A witness later identified a car belonging to ‘J’ as identical to one he saw driving toward Tommy Donaghy’s workplace minutes before the shooting itself. 131 The files found in ‘J’s possession included details on Tommy Donaghy’s family and that of Sinn Fein councillor Bernard O’Hagan – shot dead a month later. Intelligence suggested ‘J’ drove the getaway car when Bernard O’Hagan was killed. 132 Questioned about both killings, and again after the shooting dead of Danny Cassidy in April 1992, ‘J’ was never charged or convicted for involvement in any murder. 133 He was later suspected of involvement (alongside ‘Person K’) in both the Castlerock and Greysteel massacres.
‘J’ should not be seen as an isolated case. While having no jurisdiction over the military, the PONI found that a ‘significant number’ (some military documents named twenty-eight) of serving and former members of the UDR ‘had links’ with loyalists in the north west in the late 1980s and early 1990s. 134 The North West UDA/UFF were ‘receiving information from a number of former or serving members of the military’. 135 Loyalist ‘infiltration’ of the UDR at this time allowed ‘access to weapons, training, intelligence and uniforms’. In turn this ‘added to their effectiveness in carrying out sectarian attacks’. 136 Some former UDR members were also ‘senior figures within the North West UDA/UFF’. 137 One, ‘Person A’, who occupied a ‘senior position’, was suspected of being involved in importing weapons into the North alongside a then serving UDR member – the same ‘Person V’ suspected of supplying the strategic and sensitive intelligence used in UDA/UFF attacks – such as that of Gerard Casey. 138 ‘A’ led the UFF unit which killed Eddie Fullerton and was linked to several other attacks. 139
Undermining police investigations
In the aftermath of an attack, collusion can involve hampering investigations and preventing or deterring prosecutions. This has formed a fundamental concern for many victims’ families. Earlier reports provided stark evidence of what this could involve. For example, following the Loughinisland pub shooting, key suspects identified in the immediate aftermath (including a former member of the UDR) were not arrested or questioned until weeks, months and sometimes years later. 140 This severely undermined the investigation, one of a catalogue of catastrophic police failings. Loughinisland also illustrates the variety of motives this form of collusion could involve. Personal motivations, local bonds and community divisions played a part. The all-consuming focus on combating the IRA ensured loyalist killings were simply not given the same priority. However, Loughinisland showed too how protecting informers was a critical factor. This needs to be understood within a wider, institutional context.
In the early 1980s policing in the North saw a fundamental shift away from the ‘prevention and detection of crime’ toward prioritising the gathering of intelligence. 141 This was driven by several factors. A policy of ‘police primacy’ saw the expansion of the intelligence infrastructure and specialist, militarised counterinsurgency units within the RUC. 142 At the same time, the use of informers and agents became the fulcrum of the ‘intelligence war’. While the FRU was in charge of agent handling for the British Army, Special Branch came to monopolise the same role within the RUC. This underpinned a significant rise in the power and influence of Special Branch within the RUC, as a ‘force within a force’. MI5-directed changes to RUC policy and practice also increasingly ensured the recruitment and protection of agents and informers would take precedence over arrests and punishing criminality. 143 As a result, preventing the exposure, arrest and conviction of informers within loyalist paramilitary groups became embedded and institutionalised in the practices of Special Branch and it generated tensions with other investigative sections of the force – notably CID.
This is what Michael Maguire meant when he referred to an ‘intelligence mindset, which placed the collection of information before the prevention and detection of crime’ as playing a pivotal role in the collusive failings evident in the Loughinisland massacre. 144 Delaying or withholding intelligence to investigations where informers were implicated in killings was a systemic practice – known in RUC Special Branch as ‘slow waltzing’. 145 Files were marked for ‘limited’ or ‘no downward dissemination’ so that while seen by senior Special Branch officers, they were not given to CID investigations. ‘Slow waltzing’ is evident in many collusion cases. Recent investigations add to our understanding here too. So, for example, in South Belfast, a long list of ‘investigative failures’ following the Sean Graham massacre have been identified, including ‘inadequate forensic, suspect, and arrest strategies; failures to adequately test and probe evidence’ and the way certain identification parades were carried out. 146 However, what was crucial in this and other cases in the area was the ‘failure’ of Special Branch to hand over intelligence to the investigating teams – something that ‘undermine[d] the effectiveness of these investigations and impeded [bringing the] perpetrators of these serious crimes to justice’. 147
In the north west the Ombudsman found that most intelligence held by Special Branch was shared with CID. However, there were significant exceptions, including not passing on information on the role of the UDR member ‘Person V’ in the killing of Gerard Casey. 148 Information on one of those suspected of being involved in killing Eddie Fullerton was withheld from the An Garda Síochána investigation. 149 More broadly, clear evidence of the intent of the UDA/UFF in the area to escalate violence directed against republicans ‘was not accompanied by a policing response proportionate to the increased risk’. 150
There were other ways in which maintaining secrecy about state covert, counterinsurgency actions could trump justice. The case of Damien Walsh is important here. When the 17-year-old was shot dead by the UDA, a specialist covert unit of the British Parachute Regiment had his workplace under surveillance. 151 The soldiers witnessed the attack − indeed they alerted the police after it happened. The UDA gunmen were still able to make their getaway. All this was confirmed by the Ombudsman report. The presence of the covert unit had, however, long been denied by the British Army and the MoD. Crucially, the original RUC investigation was not told about the surveillance unit and so was never able to interview ‘military personnel [who were] eyewitnesses to murder’. 152 This ‘deliberate decision impeded the police investigation’. 153
There are parallels here with other collusion cases. 154 For example, 76-year-old Roseann Mallon was shot dead in the family home by the UVF in East Tyrone in 1994. 155 At the time, the house was under constant British Army surveillance, including cameras linked directly to a nearby military base. Roseann’s nephews were well known republicans. Six British soldiers were also dug into hidden observation posts around the home. Again, despite the Army unit alerting the RUC, the gunmen escaped. The British soldiers were told by the Tasking and Co-Ordinating Group (TCG) running the operation to take no action. 156 In the case of Damien Walsh, the covert paratroopers were likewise ‘directed to lift off’ – to return to base – by their TCG. 157 After Roseann Mallon’s death the family uncovered the surveillance cameras. Only then did the surveillance operation come to light. The RUC investigation into the killing had not been told about the cameras or the presence of the British soldiers. Indeed, the military continued to deny its role. When the camera recordings were eventually handed over to the investigation, those for the day of the shooting had been destroyed. 158
However, maintaining covert unit secrecy was not the only factor involved in the case of Damien Walsh. As in other cases, RUC Special Branch also made a ‘deliberate decision’ not to pass vital intelligence to the team investigating his murder. It was aware that loyalists were receiving information in order to target people in West Belfast from ‘British intelligence’ and the ‘police’. 159 It also had information as to who was involved in the shooting, but waited several months before telling the murder investigators. Vital opportunities were therefore lost ‘from police policies designed to safeguard sources of information’ – informers. 160
There were cases too of more widespread problems with the way investigations were carried out in the aftermath of attacks. In the days following the Sean Graham massacre in February 1992, the RUC had an extraordinary opportunity to seize the guns used in the attack, along with other vital forensic materials, including the clothes worn by the gunmen. 161 They knew where the weapons, clothes and other items (including a walkie-talkie and ammunition) were being kept and when the UDA was planning to move them. However, three senior RUC officers decided that no search or arrests should be made. The rifle was not recovered until two weeks later. The revolver was only seized after a separate police operation in early May. 162 By then, whatever forensic opportunities the weapons might have offered to convict those who carried out the killings had been lost. As the South Belfast PONI report notes, the failure to seize the guns and clothes used in the Ormeau Road massacre, that ‘could have yielded significant forensic evidence [meant that] one of the most significant evidential opportunities . . . was gone’. 163 All of this took place as part of a joint British Army-RUC covert operation set up in the immediate aftermath of the massacre. It was also directly overseen by the Belfast TCG, which consisted of high-ranking police and military figures responsible for coordinating intelligence-led security force operations in the area.
Quite why it was decided not to seize the guns and clothes of the Ormeau Road killers cannot be known because those records are among many that have since been destroyed. 164 However, the PONI report concludes that the decisions taken were a matter of systemic policy – ‘indicative of a strategy in which police prioritised intelligence gathering and protection of sources, over the detection of serious crime and their obligation to bring the perpetrators of those crimes to justice’. 165 While this runs directly contrary to a fundamental tenet of rule-of-law policing, it is entirely in line with the extra-legal ‘grey zone’ of counterinsurgency thinking. 166 It also evidences a policy of collusion in which stopping loyalists killing people, or arresting them for doing so, was less important than pursuing other strategic goals. To add insult to considerable injury the rifle used in the Ormeau Road massacre ended up on public display in the Imperial War Museum. 167
The ‘failure’ to carry out basic, but vital, policing operations in the wake of a mass loyalist killing was not unique to the Sean Graham massacre. It was echoed, for example, in the aftermath of the 1994 Loughinisland pub shooting. In this instance the getaway car used in the attack was found within twelve hours of the shooting. It had been abandoned close to the home of a former member of the UDR previously reported as a key planner and organiser of the UVF in the area. Despite this, and while other houses in the vicinity were ‘visited’ by the RUC, this loyalist’s home (and that of another suspect living close by) were not; evidence of ‘a reluctance by the police to conduct inquiries in the areas of the suspects’ addresses’. 168 Indeed this former soldier and leading UVF suspect was not arrested or questioned until two months later – something for which the investigating Ombudsman could find no rationale. 169 He and others had also been tipped off by an RUC officer that their arrests were imminent. 170 In this case, stated the Loughinisland PONI report, ‘the failure to conduct early intelligence-led arrests . . . seriously undermined the investigation into those responsible’. 171 Not taking advantage of the ‘range of forensic and other evidential opportunities’ through making early arrests proved ‘catastrophic’ for the investigation. 172 Records of the interviews of the key suspects when they eventually did take place have also since been destroyed. After having been held in police storage ‘open to the elements’, the police also destroyed the getaway car less than a year after the shootings.
Destroying records
Collusion can also be understood as a failure to keep − or the deliberate destruction of − incriminating records and evidence. Perhaps the most high-profile and audacious example occurred even as many of the killings examined in the recent PONI reports were taking place. In January 1990, the offices of the first Stevens inquiry into collusion, sited within an ultra-secure police complex, were destroyed by fire. 173 An initial investigation found the blaze had started accidentally. Much evidence suggests otherwise, not least that security and fire alarms had been disabled and telephone lines cut beforehand. Then, and decades later, John Stevens felt the fire was never properly investigated and viewed it as a deliberate act of arson designed to destroy the documentary evidence collected by his team. An attack he believed had been carried out by other branches of the state, namely, members of British military intelligence, FRU, aided by RUC Special Branch. 174 For journalist Ian Cobain, this was but one example of a culture of secrecy, centuries in the making and with roots deep in the history of empire, that has shaped the modern British state. From Kenya to Carrickfergus, the decades- long wholesale concealment and destruction of records has been designed, by the ‘history thieves’ of officialdom, ‘to erase all trace of the darker deeds of Britain’s colonial enterprise’. 175
This was a far from unique instance of conflict records being destroyed. Certainly, the deliberate destruction of records has been an abiding feature of collusion cases. In itself understood as evidence of collusion, it has been a key feature of covering up the extent and nature of collusion and contributing to a culture of denial. So, for example, Michael Maguire found there had been the ‘deliberate destruction of important police documents’ relating to the 1987 South African arms shipment. 176 This included the RUC TCG records of the operation to supposedly recover the weapons. 177 He also found evidence that records immediately relevant to the Loughinisland massacre had been destroyed. A major store of sensitive RUC records was held at Gough barracks in Armagh. These included the interview notes of those suspected of carrying out the Loughinisland murders as well as many other collusion cases. In July 1998, mere months after the signing of the Good Friday Agreement, reports of asbestos contamination led to the wholesale destruction of these records. 178 Senior RUC officers were aware that many of these documents were highly sensitive. However, none were retained or copied. There is not even a record of exactly who authorised the destruction of these files. 179 The RUC also had a general policy of routinely destroying recovered weapons. This prevented further ballistics checks being carried out and potential links to unsolved crimes being lost. 180
Here again, patterns repeated across the North of the deliberate destruction of official records that stymied investigations, prevented transparency and accountability and ensured impunity are evident in recent reports. In the north west, the records of RUC Special Branch agent handling no longer exist. All such files have been destroyed. 181 This included information on informer recruitment, payment and for how long they were acting as paid state agents. The destruction of such records has severely limited investigations and proved particularly ‘egregious’ where handlers were working with informers believed to be involved in ‘serious sectarian crimes, including murder’. 182 The absence of either accountability controls, or of records of agent handling, in themselves constitutes collusion. 183 In West Belfast, there were no records of the RUC Tasking and Co-ordinating Group responsible for the covert operation in place at the time of Damien Walsh’s killing. Along with what the Ombudsman describes as the ‘poor recollection’ of police witnesses this made it difficult for the investigation to establish precisely what role the covert British Army unit played at the time of his death. 184
Likewise, there is an extensive catalogue of the absence or destruction of RUC records in relation to the Sean Graham Bookmakers killings. This includes the records for the decision by the RUC specialist unit to hand back the Browning pistol to a UDA informer prior to its being used in the massacre. The lack of records for such a ‘high risk’, authorised, covert strategy made ‘scrutiny, then and now, impossible’. 185 Special Branch records of the surveillance of key suspects involved in transporting the murder weapon in the 24 hours before the attack have been destroyed and were not shared with the investigative team at the time. 186 Records of the intelligence used by the RUC TCG in directing covert investigations after the massacre have also been destroyed − something the Ombudsman found ‘inexplicable’. 187 The failure to pursue the recovery of the weapons and clothes used in an attack cannot properly be investigated because the relevant records no longer exist. 188 In sum, the ‘systematic destruction’ of both RUC TCG and Special Branch records in relation to this and other South Belfast killings ‘had the effect of obstructing not only some of the murder investigations but also examination of police accountability’. 189
Handling informers
Throughout the conflict, surveillance, interrogation and the use of agents and informers were the key intelligence-gathering means employed by the state. 190 Each in turn raised issues of state deviance. 191 By the 1980s, the informers and agents were the fulcrum of the ‘intelligence war’. The FRU was in charge of agent handling for the British army, while police primacy ensured Special Branch handled informers within the RUC. Indeed, the changes to informer recruitment and handling introduced in the early 1980s gave even more power and control to Special Branch. The structure of the RUC also allowed Special Branch – the ‘force within a force’ – and agent handling a large degree of autonomy from the local policing structures. 192
It is clear that throughout the North RUC Special Branch had informers and agents in place ‘from top to bottom’ within loyalist (and republican) armed groups in the early 1990s. 193 It has often been argued that the huge increase in loyalist killings in this period was the result of a lack of informers in their ranks. In this dominant narrative, a loyalist ‘old guard’ had largely been removed and replaced by a younger generation that was simultaneously more ruthless and less susceptible to infiltration and recruitment as informers. 194 In other words, the disruption of loyalist ranks that followed the first Stevens inquiry, the uncovering of Brian Nelson’s role as a state agent and the subsequent break-up of an ‘old guard’ (many of whom had been informers) produced a gap in police knowledge of what was going on, and a power vacuum filled by this ‘young guard’. In turn, this inadvertently facilitated an escalation in loyalist violence. Inquiries into collusion, this logic implied, were bad because they interfered with intelligence work, and so led to more people being killed. It was a narrative that absolved state forces, denied accountability and blamed ‘fellow travellers’ and ‘do-gooders’ – the manipulative and the naive − for the deaths that happened. It was an argument long employed by some academics and defenders of the RUC and British Army. It was never true.
It is also a narrative echoed in the arguments of former members of Special Branch in the recent collusion reports. So, for example, several retired Special Branch officers stated that there was a lack of intelligence about the activities of the South Belfast UDA prior to the Ormeau Road massacre. 195 Only after the event, they suggest, was there a sustained effort to target and recruit loyalist informers in the area. This was not true. The ombudsman investigation found the RUC ‘already had informants at all levels within South Belfast UDA/UFF prior to the attack, ideally placed to report on the activities of the organisation’. 196 The contrary ‘belief’ supposedly held by former Special Branch officers was therefore ‘inaccurate’. Though, given their position, no-one was better placed to know such an apparent ‘belief’ was false. Similarly, RUC Special Branch had successfully recruited informers in the north west at all levels by the late 1980s. Indeed, this is the very period when the RUC had embarked on a major drive to recruit informers in the area. Everything therefore suggests the police had a greater access to the workings of the UDA at the very time when there was a significant rise in their proven ability to kill. This sits uncomfortably with the argument that the arrival of the ‘young guard’ led to a dearth of intelligence on loyalists.
To all intents and purposes there was no regulation and few (if any) rules governing the handling of informers by RUC Special Branch in the late 1980s and early 1990s. The prevailing guidance for RUC use of informers pre-dated the conflict and were never adopted by Special Branch. 197 In 1988, the RUC admitted that such guidelines could not be ‘strictly adhered to’ without undermining an ‘effective intelligence network’. 198 This mirrored the position of British Army intelligence. FRU head Colonel Gordon Kerr derided what he called such ‘armchair rules’ during the trial of his agent Brian Nelson. 199
Although the leadership of the RUC regularly raised concerns with senior politicians and civil servants over this gap between law and practice, it was not to put a halt to what Special Branch handlers were allowing – and directing – agents and informers to do. Rather, they worried their officers were potentially vulnerable to blame (and worse) if found out. For example, in a memo to the Northern Ireland Office (NIO) in January 1987 the RUC described the Home Office guidelines as ‘unrealistic’ and sought greater clarity because ‘most senior RUC officers, including the Chief Constable, knew that colleagues involved in the authorisation and management of informants felt exposed and vulnerable’. 200
As a result, and as the de Silva review into the killing of Pat Finucane had already revealed, there was anything but a lack of awareness of the issue of agent handling at the highest levels of the state throughout this period − far from it. 201 Indeed, de Silva found it was an issue raised throughout the conflict. He also details the extensive and protracted discussions of the matter between state political, military, policing and civil administrative agencies (and at the highest level) from 1987 and into the early 1990s. This was not something hidden from government. ‘It was manifestly not the case’, argued de Silva, ‘that agent-handlers were seeking to conceal the general nature of their activities from those in authority; on the contrary, they wanted the political leadership to provide a clear framework and direction’. It was an issue ‘considered extensively at Cabinet level and Government Ministers’. However, ‘Ministers continued to place a high priority on pursuing an intelligence-led approach to the terrorist threat’. ‘The result’, de Siva found, ‘was that agent handlers and their supervisors were being asked to perform a task. . . that, in some cases, could not be carried out in a way that was both effective and lawful [my italics]’. 202 There was then, and throughout this period, a ‘lack of political will’ to introduce formal changes or resolve ‘grey areas’ involving ‘the demands placed on the intelligence sector and its legal capacity to achieve its objectives’; a perspective shared by the NIO and senior legal advisers. 203
For de Silva, this may have been a ‘wilful and abject failure’ of policy making, but one that was essentially a matter of regret for all concerned. 204 However, the absence of such guidelines might be seen not as policy failure but, rather, its purpose − to open up of a space of legal obscurity and plausible deniability that facilitated, rather than hampered, counterinsurgency practice. The lack of clear laws and guidelines was the means to allow informers and agents, and their handlers, to do things that were, to all intents and purposes, contrary to the rule of law, but consistent with the end goal of counterinsurgency. In that vein de Silva might have given more weight to his own conclusion, that the ‘system [for agent handling] appears to have facilitated political deniability . . . rather than creating mechanisms for an appropriate level of political oversight’. 205
Officially, RUC informers were supposedly told they ‘must not break the law’ and should have ‘no involvement in crime . . . [including] murder’. 206 This echoed the official position of the British Army, which stated it was ‘unlawful for any person to authorise an illegal act’. 207 At least, that was the theory. At the same time being a member of a paramilitary group was often a ‘prerequisite’ for recruitment and there was ample intelligence and evidence that loyalist informers were committing crimes, including murder. 208 In the north west a number of Special Branch informers continued to be ‘actively tasked and utilised’ by their handlers ‘despite there being intelligence and evidence linking them to serious sectarian crimes, including murder’. 209 Likewise in Belfast, the ‘previous involvement in murder’ of a senior UDA figure such as William Stobie made his recruitment ‘questionable’. 210 His ‘continued involvement in serious crime’ after becoming an agent demonstrated a lack of ‘effective oversight’, while allowing him to control (and providing him with) weapons was ‘endangering the public’. Indeed, in some cases, having been involved in killings seems to have been a positive for recruitment as an informer as far as Special Branch was concerned. One member of the South Belfast UDA/UFF was recommended as an agent precisely because he had played an ‘integral part’ and ‘prominent role . . . in the planning, preparation and execution’ of ‘multiple murders’. ‘This background in numerous murders’, found the Ombudsman, ‘was the attraction for his recruitment as an informant [my italics]’. 211 In another case, the Ombudsman found that a South Belfast UDA informant only began to be involved in serious crime ‘including murder’ after they had been recruited. 212 If the guidelines proposed that informers should not ‘break the law’, the modus operandi for Special Branch was the diametric opposite.
It is also clear that Special Branch handlers were aware many informers and agents did not provide intelligence on killings, including those in which they were themselves involved. Yet this too did not end their employment as informers. Michael Maguire found that many of the Special Branch informers in senior positions in the UVF ‘were almost certainly not asked for specific information’ that might have helped in the investigation into the Loughinisland massacre. 213 Special Branch also ‘continued to engage’ with ‘sources’ believed to have been involved ‘at some level’ in the Loughinisland murders. One ‘legitimate suspect’ was an RUC Special Branch informer at the time of the attack and continued in that role ‘for a number of years’ afterwards. 214 William Stobie might have been an ‘unreliable informant’ but that did not prevent his continued employment. 215 In the north west, informers would have had ‘detailed knowledge of attacks they were involved in’ but apparently ‘chose not to share this information’. 216 They continued to work for Special Branch. One informer apparently did not share information with his handlers that ‘could have prevented murders’. He was later rehired.
How was this explained? Despite arguing that loyalist informers were ‘low lifes’ and ‘accomplished liars’, one former Special Branch officer nevertheless suggested that handlers had to ‘accept at face value what they were being told was true’. 217 In other words, that Special Branch agent handlers, trained interrogators, would not ‘probe, test and assess information’ but accepted and then continued to work with informers, over several years, who they had good reason to believe were involved in killings. This spurious logic is unconvincing in squaring the circle of Special Branch knowledge that many of those working on their behalf were directly involved in murder.
Evidencing collusion, challenging denialism
Evidencing patterns of collusion challenges the denialism that has continued to characterise much (often politically motivated) opposition to legacy investigations. We still have only an opaque and partial picture of the work of agents of British covert military and intelligence agencies. But we know it included helping to provide weapons and intelligence files to target nationalist and republican victims. We also know that what was being done was discussed by those occupying the highest offices of the state. Political, policing and military leaders knew what was happening. Collusion was everywhere, and it was systemic.
Despite the large body of evidence that decades of official and unofficial investigation and analysis has produced, substantiating collusion allegations, legacy debates have also seen the (re-)emergence of what might be termed ‘collusion denial’ in popular, political and academic discourse. Former members of the state security forces have been to the fore in challenging collusion claims, both in the courts and in the public realm. 218 For example, the Northern Ireland Retired Police Officers Association (NIRPOA) represents former members of the RUC. Led by a one-time Head of Special Branch, the NIRPOA has mounted a series of legal challenges to official findings of collusion, 219 which includes ongoing court cases in relation to the recent Ombudsman reports. 220 There have also been concerted campaigns, such as those launched, led and fostered by pro-union, northern-based newspapers such as the Belfast Newsletter. 221 High profile political, academic and security figures have featured prominently in such campaigns. They include Belfast-born, retired British Army Colonel Tim Collins, who denounced what he termed the ‘fabricated historical narrative . . . [and] lie of collusion’. 222 The Chair of the NIRPOA has similarly led the charge against what he decries as the ‘trashing of the reputation of former RUC officers amid the ‘toxic findings of collusion’. 223 These have paralleled the vociferous calls for an end to a so-called ‘witch-hunt’ against British soldiers for state killings during the conflict from prominent and powerful quarters of the rightwing establishment in Britain.
As noted, the 2022 proposal to prevent any further Troubles-related investigations or court cases is essentially the state response to these demands to halt future prosecutions of state actors. 224 If the legislation is passed, there will be no more investigations such as those that have illuminated and given official confirmation of patterns of collusion long alleged that took place in places like South Derry, Loughinisland or in South Belfast. Far from being designed to serve the interests of victims, the legacy bill will serve to enshrine impunity and deny accountability and truth for families, and particularly where the victims were killed either directly by state forces or through collusion. Denialism works not to promote the rights of victims and survivors but to preserve the secrets of the state’s ‘dirty war’. Challenging such denial by identifying the patterns of collusion that took place during the conflict in and about the North of Ireland remains the work of a human rights-based future.
Such contemporary collusion denial − whether in terms of individual (perpetrator) or institutional/state (official) denial − derives from ‘collective, societal and individual impulses to reject or re-contextualise disturbing facts’. 225 Various techniques are adopted to neutralise the ‘moral bind of law’, both before the act (making ‘wrongdoing’ possible) and after – to ‘protect’ the perpetrator from ‘self-blame and the blame of others’. 226 This includes the ‘denial of responsibility’ (‘that was not supposed to happen’), ‘denial of injury’ (‘it was not that bad’) or ‘denial of the victim’ (‘they are really in the wrong’). The ‘condemnation of the condemners’ questions the motives of critics while an ‘appeal to higher loyalties’ invokes the bonds of group-belonging as justifying that which, within wider society, appears wrong. These ‘vital lies’ preserve collective myths and provide a ‘moral vocabulary of self-exoneration’. 227 In other words, such denial acts to ease both individual and collective cognitive dissonance created by the space between what should have been done and what actually happened. It also preserves the flawed ongoing claims by state agencies and institutions to have acted within the rule of law and so may underpin the potential of state wrongdoing to be repeated.
Footnotes
Mark McGovern is Professor in Sociology at Edge Hill University, Lancashire. His research is primarily concerned with the study of political violence, conflict and post-conflict transition, particularly in the North of Ireland. He is the author of Counterinsurgency and Collusion in Northern Ireland (London: Pluto Press, 2019).