Written evidence submitted by Dr Conor McCormick, Reader in Law, School of Law, Queen’s University Belfast – (TLA0002)

House of Commons Scottish Affairs Committee inquiry into the dual role of the Lord Advocate

Introduction

1.             This inquiry raises a question of constitutional design about how the Lord Advocate’s different responsibilities should be organised and exercised.[1] The Lord Advocate heads the system of criminal prosecutions in Scotland alongside a broad mix of further responsibilities. These include advising the Scottish Government and representing it in civil litigation; executive business as one of the Scottish Ministers, though not a member of the Scottish Cabinet; responsibility for the investigation of deaths, and other public interest functions.[2] The independence required varies with the function. Reform proposals must therefore consider not only whether prosecutorial functions should be reallocated, but where each remaining function is to go in the event of a ‘split’ and how it is to be governed.

2.             This evidence addresses all four of the Committee’s main questions, drawing on research conducted since 2015.[3] Its conclusion is that the current arrangements should be retained unless and until their existing safeguards are shown to be insufficient to sustain their constitutional legitimacy. The Committee may receive evidence about public trust which warrants a different conclusion, and may in any event weigh the competing constitutional considerations differently. If separation is chosen, important questions about the reallocation of functions must be confronted.[4]

Independence, Accountability and Trust

3.             As I have argued elsewhere, law officers in the United Kingdom can usefully be evaluated through three interdependent constitutional values, namely: independence, accountability and trust.[5] Malcolm McMillan’s recent report for the Scottish Government highlighted my framework and observed that ‘[c]onsideration may be given to applying a framework such as this one to any evaluation of the current roles of the Scottish Law Officers, and of any alternative approaches to the roles’.[6] Applied to institutional design, my approach essentially prompts the designer to consider:

i.          who must be free from whose influence when exercising which functions (the independence dimension);

ii.          who must explain and justify decisions, to whom, and with what consequences (the accountability dimension); and

iii.          what gives ministers confidence in the advice they receive while also giving wider audiences confidence that prosecutions and other public interest functions are properly handled (the trust dimension).

Four Models

4.             Any reform proposal must suggest something to put in the place of that which it would remove. Naturally, options are easier to weigh up if they are set out in an intelligible order of some kind. The order I have deployed below highlights various manifestations of decreasing political accountability over prosecutorial and other public interest functions, with a correspondingly greater distance from the centre of government represented by each rung of the ladder. The four options worthy of consideration are (A) the spliced model, (B) the superintendence model, (C) the sponsorship model, and (D) the siloed model. These are just terms of convenience for the purposes of this submission and the ladder is simply a metaphor for comparing the trade-offs each arrangement makes between the constitutional values of independence, accountability and trust. They are intended to clarify the choices facing the Committee rather than to provide a comprehensive comparative analysis.

(A) The Spliced Model

5.             The Lord Advocate’s present combination of functions illustrates the spliced model. I will begin by considering the safeguards surrounding this arrangement and the criticisms it has attracted before turning to Operation Branchform and possible improvements.

6.             Under the current arrangement, a single office-holder heads the prosecution service while carrying out the further functions outlined above. As such, there are a range of statutory and non-statutory safeguards in place which seek to ensure appropriate levels of independence, accountability and trust at any given time. For instance, appointment and removal require the Scottish Parliament’s agreement under section 48(1) of the Scotland Act 1998, and section 48(5) requires decisions as head of the systems of criminal prosecution and death investigation to be taken ‘independently of any other person’. The wider category of ‘retained functions’ is excluded from collective responsibility by section 52(5)(b). Under section 52(6), this category encompasses functions exercisable immediately before the Lord Advocate ceased to be a Minister of the Crown by virtue of devolution, together with statutory functions subsequently conferred on the office of Lord Advocate alone. When exercising such functions, the Lord Advocate acts independently of other Ministers and is not bound to abide by a collective government position. By dint of section 27(3), the Lord Advocate may also decline to answer questions in the Scottish Parliament about a particular prosecution where answering might prejudice criminal proceedings or would otherwise be contrary to the public interest. The Scottish Ministerial Code reproduces the collective responsibility exception and further confirms that while the Scottish law officers ‘act wholly independently of the Government’ in criminal proceedings, in civil proceedings ‘a distinction is to be drawn between proceedings in which the Law Officers are involved in a representative capacity on behalf of the Government, and action undertaken by them in the exercise of their various retained functions or on behalf of the general community to enforce the law as an end in itself’.[7] There is accordingly no lawful or constitutional route by which a ministerial colleague of the Lord Advocate can direct or exert improper pressure in relation to a prosecutorial decision, nor indeed various other public interest functions. Moreover, recent appointments have been made on an overtly non-party political basis, though this does not alter the office-holder’s statutory status as a Scottish Minister and thus a member of the Scottish Government.[8]

7.             An important argument in favour of the current arrangement concerns democratic responsiveness. The head of the prosecution service can explain directly to other Scottish Ministers how their decisions affect the service and press for the support it needs. Indeed, in defending the Lord Advocate’s direct voice within government, Lady Elish Angiolini, a former Lord Advocate, has emphasised that independence is ‘not just about decision making in individual cases’; it is also about ‘the resources and structures to make those decisions independently’, such that any ‘reliance or dependency on another Minister to bid for those resources may be construed as likely to diminish rather than enhance the independence of the office’.[9] The status quo also secures the prosecution head’s direct accountability to a democratically elected legislature. In this respect, accountability is more direct than where one office-holder answers for the performance of another over whom they have limited control.

8.             Most objections to the current arrangement are grounded in a competing value, namely the trust which the public must be able to place in prosecutorial decisions when those responsible for them work so closely with party political Ministers. If a politician is under investigation, or another conflict of interest arises, the established response is recusal, which means the Lord Advocate stands aside and leaves the decision in question to prosecutors and to Crown Counsel, namely the Advocates Depute who take decisions in the most serious cases. In such a context, Jim Gallagher has questioned whether it is satisfactory that the chief prosecutor should have to stand aside from precisely the cases in which public attention is greatest.[10] Similarly, Liam Kerr MSP has queried the difficulty of obtaining answers in the Scottish Parliament when both of the Scottish law officers have withdrawn and neither can speak to the case in question, among other things.[11]

9.             Operation Branchform has brought these questions of constitutional legitimacy into the spotlight over recent months.[12] A police investigation into Scottish National Party finances led to the prosecution of the party’s former chief executive, Peter Murrell, who was sentenced in June 2026.[13] The Scottish law officers were recused, and decisions about the investigation and prosecution were instead taken by prosecutors and Crown Counsel.[14] In January 2026 the then Lord Advocate nonetheless sent the First Minister a minute confirming the indictment, explaining the law officers’ exclusion and warning against prejudicial comment.[15] Critics argued that details not yet publicly reported gave an advantage to the First Minister, who also led the party whose finances were being investigated. The then Lord Advocate replied that the indictment had been served and could be made public, and that she had acted as head of the independent prosecution service.[16]

10.         This episode illustrates the importance of clear rules about who may receive information following recusal and for what purpose. The two statements subsequently published by the Crown Office and Procurator Fiscal Service in August 2026 are a creditable response to this point. The first records a convention predating devolution under which the Scottish law officers do not select the Advocate Depute who decides in such cases; instead, they ‘will be advised of the nature and scope of the case but will not be given access to the evidence’.[17] The second formalises longstanding information-sharing arrangements, requiring a lawful, necessary and proportionate disclosure, recorded reasons, and authorisation by the Crown Agent where the law officers are recused.[18]

11.         On the Committee’s particular question about Operation Branchform, a limited notification necessary to prevent prejudicial comment by Scottish Ministers is entirely defensible in principle. The potential difficulty lies in the route by which it is communicated. One option would be for a senior prosecutor rather than a recused law officer to communicate any necessary notification, subject to the authorisation and recording requirements described above.[19]

12.         Alongside that possibility, two further steps may be worth considering.

13.         The first is gathering better evidence about the extent to which public trust or distrust attaches to the present combination of functions. The published record of recent controversies certainly establishes that the current arrangement has been criticised on this basis and that its appearance troubles serious observers. But these criticisms do not necessarily establish that public trust has declined to an extent warranting structural reform, nor that a particular office-holder has acted unconstitutionally. Recent Scottish law officers have not faced votes of no confidence in the Scottish Parliament, for instance, though the absence of such votes offers only a limited indication of parliamentary confidence and cannot establish the extent of wider public trust. I am not aware of any representative evidence establishing the extent of public trust in current arrangements. Elected representatives’ experience and the evidence submitted to this inquiry will provide valuable insights, of course, but representative research could help to distinguish the extent of wider public perceptions from those of the people who choose to participate in inquiries of this kind for one reason or another. Independent polling and/or focus groups might be instructive in this respect, though questions about reform would need to be neutrally framed and accompanied by a balanced explanation of the existing arrangements and the alternatives, rather than simply inviting support for greater independence without acknowledging the trade-offs involved.

14.         The second is to require focused independent examinations of the arrangements for information sharing when law officers stand aside. His Majesty’s Inspectorate of Prosecution in Scotland could perhaps undertake such examinations, with associated reports to be published timeously in so far as the protection of confidential case material allows.[20] Such measures could help Parliament and the public to assess developments systematically whenever the Scottish law officers recuse themselves.

(B) The Superintendence Model

15.         Superintendence, as the term is used here, means moving prosecutorial functions into a separate institution while retaining a significant relationship of political accountability to a law officer. The office left behind need not be purely advisory. In England and Wales the Attorney General retains a range of ministerial and public interest functions alongside the giving of advice, and any proposals for a Scottish equivalent would have to determine which of the Lord Advocate’s present functions would move and which would stay. What follows assumes the narrower form of separation, through which only the prosecutorial functions would move.

16.         The arrangements in England and Wales merit close attention in that, perhaps counter-intuitively, superintendence does not entail unrestricted control. The Director of Public Prosecutions leads the Crown Prosecution Service (CPS) under the Attorney General’s statutory superintendence while exercising independence as regards individual case decisions.[21] A published framework agreement sets out how that relationship operates in practice, including ministerial oversight of the service in tandem with various limits on the Attorney’s involvement in casework. Consent decisions are taken independently of government, and the Attorney is excluded from cases involving parliamentarians, ministers, parties or elections unless a decision is required from them by law. Wider ministerial representations may be sought in exceptional cases to identify relevant public interest considerations, but final decisions remain with the prosecutor.[22]

17.         Accountability thus runs along two lines under this model. Under the framework agreement, the Attorney General holds the Director to account for running the service and answers to Parliament for superintendence and sponsorship. The Director must seek the law officers’ views on sensitive prosecution guidance and give particular weight to the Attorney General’s views where that guidance involves difficult judgements about competing public interests. The Director also gives evidence to committees and is the Accounting Officer for a service with its own funding vote.[23]

18.         If Scotland adopted a model of this sort, the statutory basis for the independence of the office in charge of prosecutions would require careful attention. For instance, as indicated above, the Lord Advocate’s decisions as head of the systems of criminal prosecution and death investigation must be taken independently of any other person. The Director of Public Prosecutions in England and Wales lacks an equivalent express statutory guarantee, although their independence in individual case decisions is recognised by the framework agreement. It would also be difficult to reconcile the notion of superintendence with the very strong form of independence signified by the statutory protections presently afforded to the Lord Advocate.

19.         Separation also requires arrangements that guard against improper interference while enabling prosecutors to obtain legitimate assistance from a law officer in appropriate circumstances. In 2025 the Joint Committee on the National Security Strategy examined the discontinuance of a prosecution in England and Wales concerning alleged espionage for China. It found no evidence of improper political influence and concluded that the Attorney General had acted with constitutional propriety. It nevertheless regretted missed opportunities to inform him more fully and to seek his assistance in obtaining evidence from government departments.[24]

20.         Measured by way of independence, accountability and trust metrics, the superintendence model may strengthen public trust in prosecutorial independence by dividing certain functions between two offices akin to the Attorney General and Director of Public Prosecutions for England and Wales. That separation may make political accountability less direct, however, dividing responsibility between the law officer and the Director. Clear responsibilities and effective communication would be crucial in order for the change to sustain public trust.

(C) The Sponsorship Model

21.         Sponsorship denotes a more limited relationship. The sponsor has no power to direct prosecution decisions or to supervise prosecution policy, but supports the service and answers politically for the support provided. The service supplies the information needed for that role. A law officer can help protect prosecutorial independence by securing adequate resources and representing the service’s interests within government and Parliament.[25] The trade-off is even less political accountability for the service itself than under superintendence. Greater distance from ministers may strengthen public trust in prosecution decisions, but confidence may also suffer if important questions about the service cannot be answered politically.

22.         A proposal resembling this model appeared in a Bill developed by the Northern Ireland Parliament before its prorogation in 1972. It would have created a Director of Public Prosecutions who was neither superintended nor subject to direction, but ‘responsible to the Attorney General for the due performance of the functions’ vested in their office.[26] However, the proposal in question did not proceed in its original form. The Criminal Justice Review in 2000 later recommended a ‘consultative’ relationship between a locally appointed Attorney General and the prosecution service. Those recommendations informed the arrangements enacted in the Justice (Northern Ireland) Act 2002 and brought into operation along with the devolution of justice in 2010.[27] As discussed below, the resulting relationship provides a very narrow basis for accountability through the Attorney General.

(D) The Siloed Model

23.         The siloed model describes arrangements in which prosecutorial and other public interest functions are substantially insulated from any involvement with political office-holders. Political accountability is correspondingly limited, although legal scrutiny, financial oversight and public reporting mechanisms may remain. The question is whether these other forms of accountability can compensate for the loss of a politically answerable office-holder with responsibility for the service.

24.         Northern Ireland illustrates this degree of insulation, though its arrangements include consultation and scrutiny measures. Both the Attorney General for Northern Ireland and the Director of Public Prosecutions are required by statute to exercise their functions independently of any other person.[28] The Attorney General is not a member of the Northern Ireland Executive and has little more than a consultative relationship with the Director.[29] She has no superintendence over the Director nor any power to direct prosecution decisions. As mentioned above, these arrangements followed the Criminal Justice Review required by the Belfast (Good Friday) Agreement. They must be understood against the background of controversy over criminal justice during the Northern Ireland conflict and the constitutionally pressing need to establish confidence across communities.[30]

25.         Before the devolution of justice, John Larkin KC, who was subsequently appointed to the office of Attorney General for Northern Ireland, predicted some difficulties associated with being ‘accountable for something for which he is not ultimately responsible’.[31] On the other hand, responding to the Department of Justice’s 2012 consultation on its governance and accountability, the Public Prosecution Service (PPS) has argued that enabling the Attorney General to answer Assembly questions does not require superintendence.[32] In practice, however, the Attorney General has no route to answer such questions on the floor of the Assembly. Section 25 of the Justice (Northern Ireland) Act 2002 permits participation only to the extent allowed by Standing Orders, but no such provisions have actually been made.[33]

26.         Financial scrutiny, inspection and complaints procedures provide some accountability, alongside judicial review.[34] Under the siloed model, greater independence from political office-holders is intended to strengthen public trust. The absence of a politically responsible advocate for the service may, however, make it harder to explain its work and obtain support. In 2025, Criminal Justice Inspection Northern Ireland recorded the Director of Public Prosecutions’ view that PPS concerns about the timeliness of decisions and pressures on staff were ‘unlikely to find recognition in political considerations’ and that ‘other organisations within the criminal justice system were better able to utilise political influence’.[35] Moreover, in the 2025 Northern Ireland Life and Times Survey, 45.3% of those who had heard of the PPS and expressed a view were either not very or not at all confident that it took prosecution decisions independently.[36] This suggests that there exists substantial public doubt about prosecutorial independence in Northern Ireland, though of course those doubts could be greater still in the absence of a siloed constitutional structure. Expectations of political answerability are likely to be stronger in Scotland, however, where the Lord Advocate has long answered to Parliament for the prosecution service.[37] Removing that familiar route to accountability could itself weaken trust.

Comparison of the Four Models

27.         The following table summarises the potential strengths and trade-offs identified above.
 

28.         This comparative overview focuses on prosecutorial functions and does not address the allocation of the Lord Advocate’s other responsibilities. The consequent effects on public trust would depend on the safeguards and practical operation of each arrangement.

Conclusions

29.         The Committee’s call for evidence poses four main questions. I would summarise my direct response to each of them as follows.

i.          The chief advantage of the spliced model is that it ties the head of prosecutions directly to an elected legislature, and gives that office-holder a voice within government when resources and legislation are required. The chief drawback is the risk to public trust when the combination of roles calls the appearance of independence into question. But it must be stressed that the existing safeguards are substantial. Decisions as head of the systems of criminal prosecution and death investigation are protected by an express statutory requirement of independence, while the wider retained functions of the office are excluded from collective responsibility. In the absence of representative evidence to suggest that public distrust warrants it, or something of a similar order, I do not think the current combination of functions needs to be split between different office-holders.

ii.          In relation to Operation Branchform and the handling of conflicts of interest, confirming a law officer’s withdrawal from case decisions does not by itself appear to fully answer recently expressed concerns about the extent of their role in such circumstances. The Crown Office’s August 2026 statements provide welcome clarification, but in addition it would make sense for a senior prosecutor rather than a recused law officer to communicate any necessary notifications about relevant prosecutions to Scottish Ministers, subject to pre-existing authorisation and recording requirements.

iii.          As to how senior prosecutors should be accountable, and to whom, the Scottish Parliament appears to remain the most appropriate forum for routine scrutiny. The Lord Advocate should continue to answer to it for the performance of the prosecution service, subject to well-established limits on the discussion of individual cases. Other Scottish Ministers should answer for their responsibilities concerning funding and legislation. The UK Parliament’s principal contributions should continue to focus on reserved matters including the overarching constitutional framework. Significant reforms to the office of Lord Advocate would engage that role, of course, given that the Lord Advocate’s position as head of the systems of criminal prosecution and death investigation is entrenched by the Scotland Act 1998.[38] It should be noted here that the independent review announced by the First Minister of Scotland on 29 September 2026, chaired by Lord Boyd of Duncansby, will examine the roles of both Scottish law officers and report to Scottish Ministers in 2027.[39] Its terms expressly leave open the possibility of retaining the present arrangements, but any proposals for structural change arising from the review would need to be considered in light of the respective legislative responsibilities of the two Parliaments.

iv.          There remains a case to be made for a political law officer model as regards the perceived trustworthiness of governmental legal advice. Political standing can help secure ministers’ trust and ensure that legal advice is heeded.[40] A related consideration is the political attention given to justice issues, which actually prompted the Law Society of England and Wales in 2024 to call for a return to politically active Counsel General appointments.[41] Party discipline must nevertheless yield whenever an office-holder is required to exercise independent judgment, including in prosecution and other public interest decisions.[42] The modern-day acceptability of this model in Scotland is unclear. The recent practice of non-party political appointments may be helping to sustain public trust and confidence in the operation of the spliced model.

30.         Should the Committee take a different view and consider a continuation of the spliced model unsustainable for any reason, then the most useful aspects of this submission may be its comparative content rather than its tentative conclusions. Moreover, I recognise that both the Committee’s inquiry and Lord Boyd’s forthcoming review for the Scottish Government will be able to test the analysis offered here against a wider body of evidence and explore the practical implications of the different models in greater depth. I would simply caution that each alternative to the status quo can only achieve something by giving something up. Any reform proposal adopted without reckoning with what it would cost risks substituting one set of legitimacy problems for another.

30 September 2026