Out of Hours:
07710 077 419 
020 7520 6000  
East Anglia:
01245 280 880

SFO v Barclays: Elusive corporate criminal liability in the UK

3. Dec 2020

By Helena Spector [1]

In December 2018 the attempt by the Serious Fraud Office (SFO) to prosecute the Barclays companies[2] for crimes of corporate dishonesty failed in limine because the courts held that the dishonest conduct it alleged could not be attributed to the companies.[3] This article discusses the implications of this ruling.


The facts of the case related to the period immediately after the 2008 financial crisis, in which Barclays engaged in consecutive capital raising exercises to avoid a bailout by the UK government.

In June 2008, a first capital raising (CR1) secured a £4.4 billion investment from Qatari entities in exchange for shares issued at an agreed discount, plus a commission. This was set out in the public prospectus as 1.5%. Later a second and similar exercise in capital raising (CR2) later brought in a further £6.8 billion. This time Barclays publicly announced that the Qatari investors would variously receive between 2% and 4% commission in addition to an Arrangement Fee of £66 million. Further, on 8 October 2008, a Qatari loan to Barclays of US $2 billion was agreed which was later raised, on 29 October 2008, to US $3 billion.

The SFO alleged that the financial arrangements between Barclays and Qatar significantly differed from those publicly announced, authorised and warranted by the Bank in its prospectuses and subscription agreements. It alleged that Barclays had secretly paid much more by means of two subsidiary agreements called “Advisory Service Agreements” (ASAs). These were sham devices designed by the Barclays CEO (JV) and Group Finance Director (CL) Other individuals implicated were the Barclays Capital Executive Chairman of Investment Management in the Middle East and North Africa (RJ) and the Barclays Wealth Management Chief Executive Officer (TK). The SFO’s case was that the Barclays Board of Directors and other relevant committees within the bank, including the sub-committee responsible for overseeing offers made (the BFC), were kept in the dark about the true intent behind the ASAs and the loan.

Over these transactions the SFO brought against the bank four charges: two counts of conspiring to commit offences of fraud contrary to s.2 of the Fraud Act 2006, and two counts of giving of unlawful financial assistance for the acquisition of its shares, contrary to s.151 of the Companies Act 1985. In response to this, the bank made a preliminary application to Jay J to dismiss the charges,[4]which he did.[5]The SFO then applied to prefer a voluntary bill of indictment against Barclays for the fraudulent actions of JV and CL.[6]The application was heard before Davis LJ, who concluded that Jay J had reached the right decision on the dismissal application: on the constitution of Barclays and on the facts of CR1 and CR2 negotiations, there was no way that the potential dishonesty on the part of JV and CL could be said to extend to the Barclays Board of Directors, and therefore extend to the company as a whole.[7]

What is the position now for UK corporate liability for criminal offences?

In the wake of Barclays, it appears to be the case that in order for a company to be prosecuted for fraud or other offences that do not impose strict liability, the individual company agents conducting the wrongdoing must either be its “directing mind and will” for all purposes, or the directing mind and will for the purpose of performing the particular function in question.

Prima facie, this appears to be an impossibly high bar to achieve in anything other than a very small company. This is precisely because the organisational structure of almost all modern companies is designed to prevent any single individual from simultaneously having full discretion to act and no accountability to any others.

That the bar was raised in establishing the identification principle can be seen in the differentiation which Davis LJ made between Barclays and the civil case of El Ajou.[8] The issue in that case was whether the defendant company (DLH) was constituted a constructive trustee of large sums received and disbursed on the grounds of knowledge that such money represented the proceeds of a fraud. The board of DLH had no such knowledge, although the non-executive chairman (F) concerned in the receipt and disbursement of the money on its behalf did.

Davis LJ held that corporate liability could be established in El Ajou but not in Barclays as in the former the individual with the dishonest mens rea not only had entire control over negotiations but had also been permitted by the board of the company to exercise such control autonomously. Concurrently, it was held that there could be no defence to the proceedings that the board of directors had not known or authorised the non-executive chairman’s unlawful activity precisely because it had delegated control of the entirety of the transaction.

Importantly, Davis LJ recognised that the board of directors in El Ajou had not given the non-executive chairman entire control by formal resolution, but rather that such control had been de facto surrendered and transferred. In theory, this leaves room in future cases to establish the identification principle based on the realities of a complex, delegated business environment.

Unfortunately, this is where Davis LJ’s reasoning became slightly less cogent. He distinguished Barclays on the ground that the various resolutions of its Board of Directors indicated “the [limited] level of delegation sanctioned by the appropriate organs of the company.” Dismissing claims to distinguish form from substance in accordance with notions of effective autonomy, Davis LJ remarked that in this case “the form is the substance.” This sits uneasily with his earlier acceptance of the de facto decision making in El Ajou, in which the company did have a formal constitution ascribing overall control of the relevant transactions to motions passed by the board of directors rather than the relevant chairman.

As a matter of logic, it is hard to see how a theoretical distinction between the de jure and de facto can be maintained if Davis LJ was willing to accept that the de jure decision making structures can represent and legally constitute substantive control. Either the distinction is theoretically and evidentially material, or it is not. Under Davis LJ’s approach,

corporations can evade prosecution by simply evidencing that the board of directors retained ultimate control and/or exercised intermittent scrutiny, despite the fact that company executives were able to operate dishonestly around and possibly by virtue of these ultimately ineffective frameworks.

Indeed, Davis LJ points to the fact that the BFC could have prevented CR1 and CR2 from proceeding to conclusion, revealing the limits of executive control. It is hard if not impossible to imagine a scenario where a Board of Directors would wholesale empower a CEO to enter into significant financial contracts with investors without retaining any kind of power of veto, or at the very least control over the pro- cess by which any decision would come to be made (i.e. controlling who constitutes subordinate boards empowered to make such decisions). Barclays therefore sets a very high bar for prosecutors to prove corporate criminal liability, requiring the dishonest agents to have no less than “entire autonomy” over a deal.

[T]hat the individuals had some degree of autonomy is not enough. It had to be shown, if criminal culpability was capable of being attributed to Barclays, that they had entire autonomy to do the deal in question [122].

How can corporate liability for agent wrongdoing be established following Barclays?

The options open to bodies looking to prosecute corporations for fraudulent activities appear significantly narrowed following Barclays, in both a legal and evidential sense. This is a decision of the High Court and it remains to be seen whether in future the point is ever litigated at the Court of Appeal or Supreme Court. However, the decision was handed down by a Lord Justice of Appeal, unusually for an application to lodge a voluntary bill of indictment, which in- creases the likelihood that the decision will be met with a judicial consensus.

How then might a corporation be held liable for the fraud committed by its agents in its name? One remaining method would be through reserving charges of agent dishonesty for cases with a different factual nexus of agent autonomy. It does not seem from Barclays that this could only arise where the relevant agent is fully authorised to “do the deal” as in El Ajou, which is likely to be difficult. Instead, there is evidence from Davis LJ’s judgment that it might be more conducive to shift the evidential focus from the CEO’s or other officer’s formal independence towards the factual input, oversight and interventionism of any given board of directors. Despite the difficulties with Davis LJ’s reasoning, he does seemingly base his argument on an evidential distinction between rubber stamp exercises by a board of directors and their effective control.

For example, in support of his conclusion that the officers were not independent, Davis LJ places weight on the numerous resolutions passed by the Barclay’s Board of Directors, including the decision to delegate authority to a sub-committee to oversee the placing of offers on behalf of the Board (i.e. the BFC), BFC meetings approving the various terms of the CR1 and CR2, the Board resolution that Barclays should pay “such fees, commissions and expenses” in connection with the Qatari subscription as seemed reasonable and the Board’s approval of fees to be paid to the key subscribers (with the exception of the further unapproved £280 million). Not only did the Board and BFC retain ultimate control, but at each stage they implemented processes designed to regulate and, where necessary, delegate authority in strictly limited and controlled capacities. Absent these findings of control exercised in fact, he might have concluded, perhaps, that the officers had been “left to do the deal”.

A second method could be through confining charges to regulatory or strict liability offences, such as those set out in the Companies Act 1985 or Bribery Act 2010. In his judgment, Davis LJ was quick to point out that the Fraud Act 2006 had not been drafted with the position of corporations in mind, as evidenced by the lack of strict liability or availability of a statutory defence that it had adequate protections in place (e.g. as per s.7 Bribery Act 2010). Accordingly, refusing to hold a company criminally dishonest by virtue of the dishonest actions of a CEO could not be said to frustrate the will of Parliament. On the other hand, under s.151 of the Companies Act 1985, a company may be criminally liable for unlawful financial assistance without needing to establish dishonesty on the part of any agent. As such, it is more evidentially straightforward to establish the actus reus of specific agents and attribute liability to the company as a whole under s.151. Indeed, Davis LJ commented that the SFO would have had a greater chance at succeeding on Counts 3 and 4, those concerned with s.151, had the charge not been particularised on the basis of the relevant agents’ dishonesty (thereby rendering the issue a moot one in Barclays).

Third, greater precision might be required in the manner in which cases against companies are particularised, especially if dishonesty is alleged. Davis LJ focussed on the fact that the particularised allegations were linked to the accuracy and truthfulness not only of the specific ASA agreements but of the Prospectuses and Subscription Agreements for CR1 and CR2 as a whole, “to be viewed realistically in the round as one transaction” [at 115]. The problem was, however, that while the relevant agents had the authority to negotiate and conclude specific arrangements, they were not authorised to complete, conclude and issue the Subscription Agreements and Prospectuses. This was always and inevitably the remit of the BFC in the first instance and the Board of Directors in the second. As such, JV and CL could not have been authorised to “do the deal” and accordingly the case failed. A tighter and less ambitious focus when drafting the indictment might enable companies to be held liable for specific dishonest transactions.

However, the problems with this approach are twofold. First, the reason the SFO focussed as it did on the wider CR1 and CR2 transactions is because, particularised in the narrower alternative, Barclays could have defended itself on the basis that the company did not gain from the ASAs themselves, as distinct from agreements CRs 1 and 2 from which gain is self-evident. This is primarily an evidential issue. It is clear that JV, CL, RB and TK were cautious of clearly linking the ASAs to a wider context of investment by way of CRs 1 and 2. For example, a telephone conversation between RB and TK in June 2008 talks of doing a “side deal” with no further elaboration and elsewhere there was stated the need to discuss the ASAs by telephone rather than email. By a similar logic, a defendant company could also argue that these were rogue subsidiary deals, and the culpable individual agents could not have been the directing mind and will of the company as these were acts which the agents had no authority to do and about which the boards of the company had no knowledge nor awareness. Second and more pragmatically, limiting fraudulent activities to ancillary transactions of lesser value will have the likely effect of significantly diminishing the size of fines payable upon conviction. If the SFO intends to prosecute such offences to act as a deterrent, a low value fine arguably negates the purpose of pursuing a conviction in the first place.

Ultimately, Davis LJ recognised that the present difficulty with establishing corporate liability for fraud for the benefit of a company is one for Parliament to fix. As the law stands, the decision in Barclays seems likely to disincentivise charges being brought against companies for fraud committed by directors, even where the evidence of CEO dishonesty is strong and the company evidently profited from the fraud.

Corporate criminal liability: a problem to be fixed?

Much of the discussion on corporate criminal liability elides a central question: why should criminal liability be attached to the legal fiction of the company? Or, to paraphrase, why should stakeholders who have no criminal culpability suffer the financial and reputational repercussions resulting from a criminal conviction? The need for such liability was alluded to by the Privy Council in Meridian[9] but was not given sustained attention.[10] A common justification[11] is that the criminal law is preventative: the communicative function of a criminal law coveys that a criminally unlawful act is a para- mount social problem on a par with other, more traditional social wrongs elevated to the status of crimes. As such, the criminal label attached to kinds of specifically corporate wrongdoing carries a stigma sufficient to disincentivise corporate complacency. A more persuasive justification is that it is apposite: corporate structures – organisational networks, markets and supply chains – permit and enable individuals to cause certain harms (on a certain scale) that would not be possible without them.[12]Human rights violations committed in explicitly corporate supply chains or networks of fraudulent insider trading with serious implications for people’s pensions are both examples of this. The presence or absence of culpability is moot, especially given that it is well known that criminal law in England and Wales justifies the imposition of liability absent any mental element for an array of offences to provide protection and redress for social wrongs.[13] Economic crime conducted by corporate agents by virtue of the corporate context is one such wrong.

Where next for corporate criminal liability?

In light of the current difficulties with effectively regulating corporate wrongdoing, several options for legal reform

have been suggested.[14] First is the extension of the “failure to prevent” (FTP) offences set out in the Bribery Act to be applied to a wider section of economic crimes.[15] Under this model, “economic crime” would be broadly defined to include a range of offences such as fraud, theft, false accounting, forgery, destroying company documents, money laundering and those offences covered under the Financial Services and Markets Act 2000 – a list based on those to which a Deferred Prosecution Agreement can be obtained as set out in Sch.17 of the Crime and Courts Act 2013. Although such FTP offences were subject to completed consultations in 2017 and considered at committee stage they were never enacted. The strengths of this model lie in the fact that FTP offers a broader capacity than the identification principle to penetrate the diverse and diffuse spread of corporate actors. It seeks to recognise that corporate activity exists upon private individuals acting by virtue of complex corporate structures, bodies, agents, subsidiaries and intermediaries: and targets the organisational cultures which create the conditions for economic crime to take place and generate profit. The argument runs that by reversing the burden of proof (as under s.7 Bribery Act), the extension of FTP would provide a powerful method for holding corporations liable for the wrongdoing of its agents in pursuit of contractual or commercial advantage.

There are two main problems with the enforcement of FTP liability, problems which pose particular challenges to prosecuting fraud offences. The first, which similarly plagues any model premised on an extension of the principles of vicarious liability, is that it relies upon a derivative liability: it is contingent on establishing the criminality of the relevant individuals which recent caselaw has struggled to do (e.g. Barclays and the case of Sarclad[16]). It is a shortcoming that has been identified by a number of recent commentators.[17] Although an individual at a lower organisational level in a company may be directly responsible for an act, they might not have sufficient knowledge of the circumstances to ac- quire the requisite mens rea – which for fraud would be dishonesty. Whereas prosecuting an offence such as bribery under FTP is underpinned by a substantive offence – that of inducing the improper performance of a function or activity having been committed by one or more employees - the only element of fraud which distinguishes it from otherwise lawful action is dishonesty. But dishonesty in the corporate context might not always attach to an underlying offence committed by individual agents, or possibly should not attach to the individual agents. The recent mis-selling scandals point towards a “criminogenic corporate culture”[18] which cannot be reducible to decisions made by “front line” employees but also which ignore sales targets, sale policies and risk aversion strategies. There are exceptions to this, notably the conviction in Hayes[19] for mis-selling LIBOR, yet there are clearly significant obstacles with this approach: only thirteen charges were brought in total by the SFO for LIBOR-rigging, resulting in four convictions. This suggests that it remains difficult to challenge the honesty of the relevant employees where the mis-selling was encouraged as part of employment contracts, unofficial corporate targets, or perhaps consistently with an industry-wide practice. Without dishonest agents, fraud cannot be established under FTP as there is no substantive offence upon which parasitic corporate liability can attach.

The second problem is that any legislation would also need to causally link the wrongdoing of the agent to gain intended for a specific corporation. As a corollary, under s.7 of the Bribery Act 2010 the associated Guidance stresses the need to establish that the associated person was “com- mitting bribery on the organisation’s behalf”. Taking fraud again as a model, there could be an array of evidential and legal problems relating to whether (a) the wrongdoing was on behalf of any organisation and (b) what that organisation is, if the agent was a member of a subsidiary organisation. It is highly unlikely that any law would be drafted so that economic crimes committed on behalf of a subsidiary company by one of its employees or agents would automatically involve liability on the part of the parent company, or that liability could flow from subsidiary to parent company through straightforward corporate ownership or investment. Thus establishing intended benefit to a parent company could become a major stumbling block to bringing successful prosecutions under FTP.[20]

As a result of the first of these problems, that of derivative liability, Robin Lööf[21] has suggested a new model of corporate criminal liability based on causation for the criminal harms arising, regardless of whether any individual could be found criminally liable. Lööf proposes that corporations which occasion harm by means of any relevant actus reus for the “economic crime” offences would be strictly liable, subject to defences for which they would bear the legal bur- den of proof that there was no harm caused; namely that they were not a significant cause of the harm, or novus actus interveniens.

The causation model clearly has the advantage of better focussing prosecutions on instances of actual harm caused and removes the need to establish guilty intent for any relevant individuals. It certainly lifts a significant part of the evidential burden from the prosecuting body.

However, the main drawback of this model is that it would further complicate the way in which economic crime is prosecuted. Lööf clarifies that causation-based liability for economic crime is not relevant for offences such as bribery or fraud. Given that section 7 bribery is currently governed under an FTP model and fraud through the identification principle, a causation-based fraud model would lead to three starkly different approaches to criminalising corporate mis- feasance. Moreover, the causation model is incapable of addressing the fundamental problem raised by Barclays: how to prosecute corporations for the fraudulent conduct of its employees. As a matter of legal principle, the offence of fraud is complete regardless of whether the fraud was successful (as bribery is complete whether or not the bribe is accepted). The harm caused is irrelevant, so an approach based on harm caused would not catch the majority of fraud offences. As a result, a further mode of attributing liability would be required to prosecute corporate fraud.

This leaves a third option for prosecuting corporate liability, proposed by Stephen Copp and Alison Cronin: introducing an evidential presumption of a dishonest mens rea attributable to corporations charged with fraud.[22] Taking fraud by false representation, for example, an approach where dis-honesty is presumed would allow a prosecution for fraud to be brought against a corporation where there has been a false or misleading statement made with a view to making a gain or causing another to suffer a loss, where an ordinary person would consider such a representation dishonest.[23]As with all evidential presumptions, this would operate to shift the evidential burden to the defence to establish that the allegedly fraudulent conduct was honest.[24] This approach has the advantage of assuming corporate dishonesty to commit fraud without the need of a metaphysical “mind”, while not disturbing the formula of actus reus/mens rea formulating fraud as a criminal offence.

Although the evidential presumption for dishonesty would be the most effective and least complicated tool for prosecuting corporate fraud in the UK, it is clear that there is no one-size-fits-all answer. FTP is still likely to be a useful tool in the prosecution of the kinds of economic crime which involve the evidently culpable “rogue” employees and could apply effectively to offences such as theft, false accounting, forgery etc. Bringing a wider range of economic offences under the FTP model would likely permit the SFO or other bodies to indict more corporations. In terms of prosecuting corporate fraud, a mode of attributing liability through reversing the evidential burden of dishonesty appears to be the most pragmatic proposal and could embolden the SFO to bring charges of corporate fraud in the wake of Barclays. However, any reform of this area of the law requires a clear notion of the purpose of attaching criminal liability to corporations: the social impetus of tackling “corporate criminogenic culture”. On this point, political inertia rather than legal inflexibility is likely to be the biggest obstacle to effectively criminalising corporate bodies.

This article first appeared in the Archbold Review: Issue 10, on 21 December 2020. Helena Spector is a Probationary tenant at Red Lion Chambers.

[1] Probationary tenant at Red Lion Chambers.
[2] There were two Barclays defendants: Barclays plc, and Barclays Bank plc, which was added as a later defendant after Barclays plc had been charged. The application to dismiss concerned both Barclays defendants, collectively known as ”the companies”.
[3] Serious Fraud Office v Barclays plc [2018] EWHC 3055 (QB).
[4] Crime and Disorder Act 1998 Sch.3 para. 2(1).
[5] R v Barclays plc (unreported, 21 May 2018) (Southwark Crown Court).
[6] Although a prosecuting body may appeal to the Court of Appeal against a ‘terminating ruling’ once the case has gone to trial, where an application to dismiss succeeds the only possible remedy is an application apply to the High Court for a voluntary bill of indictment to try and recommence proceedings, which is only granted in exceptional circumstances. It is made explicit in the Practice Direction 10.B.4 that: “The preferment of a voluntary bill is an exceptional procedure.” The prosecution in seeking the preferment of a voluntary bill must demonstrate that the tribunal “was obviously wrong or unreasonable” to dismiss the charge (R v Davenport [2005] EWHC 2828 at [22]).
[7] Serious Fraud Office v Barclays plc [2018] EWHC 3055 (QB).
[8] El Ajou v Dollar Land Holdings plc [1994] 2 All ER 685.
[9] Meridian Global Funds v Securities Commission [1995] 2 A.C 500, [1995] UKPC 1.
[10] However, it is noted that the Meridian decision had a highly limited practical impact on the identification principle or other judicially viable modes of attributing corporate criminal liability for economic crimes.
[11] E.g. M. Dsouza, “The corporate agent in criminal law - an argument for comprehensive identification” (2020) C.L.J., 79(1), 91-119; M. Diamantis, "Corporate Criminal Minds" (2016)
91(5) Notre Dame L.Rev. 2049.
[12] E. Sutherland, “The White-Collar Criminal” (1940) 5 American Sociological Review 1.
[13] E.g. see discussion in Sweet v Parsley [1970] A.C. 132 at p.149; Kirkland v Robinson [1978] JP 3777.
[14] In this article, I focus on three contemporary suggestions: “fail to prevent”, criminal liability based on harm caused and the evidential presumption of a dishonest mens rea attributable to corporations charged with fraud. There are other possible approaches, including the compromise approach is taken in the context of corporate manslaughter; the principle of vicarious liability; and the ‘organisational’ model proposed by Celia Wells. See C. Wells, “Medical Manslaughter - Organisational Liability” in: Danielle Griffiths, Andrew Sanders (eds) Bioethics, Medicine and Criminal Law (2013). Cambridge University Press, pp. 192-209.
[15] E.g. C. Wells, “Corporate failure to prevent economic crime - a proposal” (2017) Crim. L.R.6.; S.F. Copp and A. Cronin, “New models of corporate criminality: the development and relative effectiveness of “failure to prevent” offences” (2018) Comp. Law. 39(4); A. Ashworth, “A new generation of omissions offences?” (2018) Crim. L.R.5.
[16] Serious Fraud Office v XYZ Ltd [2016] 7 WLUK 211; [2016] Lloyd's Rep. F.C. 517.
[17] R. Lööf, “Corporate agency and white collar crime – an experience-led case for causation- based corporate liability for criminal harms” (2020) Crim. L.R.4.; S.F. Copp and A. Cronin, “New models of corporate criminality: the problem of corporate fraud – prevention or cure?” (2018) Comp. Law. 39(5).
[18] S.F. Copp and A. Cronin, “New models of corporate criminality” p.146.
[19] R v Hayes [2015] EWCA 1944. In that case, an individual was found guilty of conspiracy to defraud the LIBOR rate, as his behaviour was deemed dishonest according to the objective standard, i.e. the ordinary standards of reasonable and honest people.
[20] Prosecutions could still be brought against the subsidiary company, but this would not allow the parent, high-value companies to be held to account.
[21] R. Lööf, “Corporate agency and white-collar crime – an experience-led case for causation- based corporate liability for criminal harms” (2020) Crim. L.R.4.
[22] S.F. Copp and A. Cronin, “New models of corporate criminality.”
[23] In line with the Supreme Court decision in Ivey v Genting Casinos UK Ltd [2017] UKSC 67 as interpreted in Barton and Booth [2020] EWCA Crim 575.
[24] It is conceivable that a straightforward amendment to the Criminal Practice Rules and clarified with a Criminal Practice Direction or guidance might suffice to enforce this proposal rather than a wholesale statutory change.