Deferred prosecution agreements: lessons from Serco

At the end of April this year, the Serious Fraud Office’s prosecution of two former Serco executives dramatically collapsed after it emerged the agency had made mistakes when disclosing documents that jeopardised the trial.

In February 2014, deferred prosecution agreements (DPAs) were introduced in England and Wales under section 45 and schedule 17 to the Crime and Courts Act 2013. They were designed as a tool to enable the Serious Fraud Office (SFO) and the Crown Prosecution Service (CPS) to reach agreements with corporate organisations to defer their prosecution for a criminal offence, in exchange for certain conditions being met.

Efficient justice

The rationale underpinning the DPA regime, as stated by the government in its 2012 response to the consultation on DPAs, is to “allow prosecutors to hold offending organisations to account for their wrongdoing in a focused way without the uncertainty, expense, complexity or length of a criminal trial”.

Furthermore, DPAs would enable prosecutors to bring more cases to justice, and secure outcomes, including restitution for victims, more quickly and efficiently.

Since February 2014, nine DPAs have been entered into between the SFO and corporates, all approved, as required in law, by the court as being “in the interests of justice” and containing terms that are “fair, reasonable and proportionate”.

Each DPA has contained a “statement of facts” (SoF) agreed between the SFO and the corporate, which details the facts underpinning the criminality for which the corporate defendant has accepted responsibility.

However, the SFO has failed to secure a single conviction against any of the individuals it has subsequently prosecuted in connection with the conduct underpinning those DPAs.

The ability of the DPA process to readily secure the acceptance of criminality by a corporate entity, on evidence which does not survive the robust scrutiny of a criminal prosecution, raises important questions: is the DPA regime meeting its stated objectives; and even if so, to what degree of success?

Serco Geografix and beyond

The stark contrast between the SFO’s ability to secure lucrative DPAs with corporate suspects and its inability to secure convictions of the individuals whose alleged conduct underlies these DPAs, is illuminated by the recent collapse of the SFO’s trial against two former directors of Serco Geografix Limited (SGL).

In the DPA between the SFO and SGL, entered into in July 2019, SGL accepted responsibility for fraud and false accounting based on the actions of two former directors. It agreed to pay a financial penalty of £19.2m and SFO costs of £3.7m.

Both ex-SGL directors were subsequently charged with fraud but four weeks into their trial, the SFO (in its own words) “uncovered errors made in the non-disclosure of certain materials”.

The judge considered these errors to be extremely serious, refusing the SFO’s application for an adjournment to allow it to remedy the position and facilitate a retrial.

The judge further commented: “It seems to me there are… real concerns in relation to the nature of the prosecution case against these defendants.”  This left the SFO in the unfortunate but self-inflicted position of having to offer no evidence.

Compounding its woes, the SFO is likely to have to pay significant defence costs. It comes as no surprise that the SFO stated that it is “considering how best to undertake an assessment to prevent this from happening in the future”.

The SFO has had some success prosecuting corporates. It successfully prosecuted Smith and Ouzman Ltd for an offence of corruption (the first conviction of a corporate for overseas bribery) in 2014, secured a guilty plea to an offence of failing to prevent bribery by Sweett Group plc in December 2015.

More recently in April 2021, it secured a guilty plea to one count of corruption by GPT Special Project Management Ltd.

However, it is the DPA regime which has delivered most of the SFO’s success against corporates. To date, it has secured nine DPAs resulting in over £1.5bn in financial penalties and in excess of £32m in costs.

This is significant when one recalls that the SFO’s core funding for the year 2019-2020 was approximately £52.5m.

But despite this success, the SFO has failed to secure a conviction against any of the 11 individuals it has prosecuted in the four prosecutions that have resulted from a DPA: Sarclad Ltd, Tesco Stores Ltd, Güralp Systems Ltd and now SGL.

Does this anomaly demonstrate a systemic problem in the DPA regime, in the SFO’s ability to prepare and bring a case to trial – or both?

Systemic problems

When considering whether to enter into a DPA with a corporate, in accordance with paragraph 1.2 of the DPA Code of Practice (the Code), prosecutors must satisfy themselves that there is evidential sufficiency and that it is in the public interest to proceed.

In considering the evidential stage, they must be satisfied that there is:

• sufficient evidence upon which a properly directed jury is more likely than not to convict (ie the ordinary evidential test that each prosecutor must be satisfied of in any prosecution); or if not, then
• a reasonable suspicion based upon some admissible evidence that the corporate entity has committed the offence, and that there are reasonable grounds for believing that a continued investigation would provide further admissible evidence within a reasonable period of time, so that all the evidence taken together would be capable of establishing a realistic prospect of conviction.

The second limb of the evidential stage is notably more diluted than the first and requires a degree of tasseography.

In essence, the prosecutor can have a reasonable suspicion that an offence has been committed, but they must also believe that continued investigation would yield admissible evidence within a ‘reasonable period of time’ that would satisfy the evidential sufficiency test.

It is worth noting that of the nine DPAs agreed to date, it is clear that four (Sarclad, Tesco, Güralp and G4S) have proceeded under the first limb of the evidential stage and three have proceeded under the second limb (Standard Bank, Airbus, and SGL).

Considering that it took the SFO six years from commencing its SGL investigation in 2013 to obtain a DPA, it is remarkable that despite such a lengthy period of investigation, the SFO was still unable to satisfy itself of the ordinary evidential stage test.

Given that SGL was a domestic case and did not require any mutual legal assistance requests, it is unclear what further evidence the SFO believed a continued investigation would yield.

This leads to the next question: does the existence of the second limb encourage speculation by prosecutors at the key evidential stage?

A false sense of accuracy

Recent commentary on DPAs has focused on a few key themes: an alleged lack of judicial scrutiny in the approval hearings (apparently evident from the swiftness of the process); the fact that judgments have become progressively shorter; and the non-publication of hearing transcripts.

Perhaps the real question is, ‘what has been scrutinised?’ – to which the answer appears to be ‘whether the company should be avoiding prosecution or paying a bigger financial penalty given the agreed facts’.

There has been no judicial consideration at all of whether the SoF is justified by the evidence and, in particular, whether the evidence justified an agreed ‘fact’ that an individual (who is not a party to the DPA proceedings) had committed a criminal offence.

While a DPA does require judicial approval, the judge does not (and is not required to) assess whether the evidence underpinning it is sufficient to establish the agreed facts; the judge’s task is instead to assess and declare that the DPA is in the interests of justice and that its terms are fair, reasonable and proportionate.

Understandably, judges will not be encouraged by either the SFO or the relevant corporate to test the evidence underpinning the agreed facts. Rather, their focus will be on having the DPA judicially approved.

In all these circumstances, it is hardly surprising that:

• In many cases where the SFO has subsequently properly considered the evidence, no individual was prosecuted and
• In other cases where the SFO has prosecuted individuals thus far, they have been subsequently acquitted.

Most lawyers in the UK are of the view that our adversarial system produces a high quality of justice, involving a robust assessment of all the relevant evidence by a court.

By marked contrast, the DPA regime is consensual and not adversarial and does not require judicial scrutiny of the underlying evidence.

It can also be said that it encourages the SFO to become wedded to a view of the evidence that it thinks has been approved by the court.

It is interesting that in response to an interview in the Daily Telegraph by one of the acquitted SGL defendants, the SFO was quoted as saying that the SoF was approved by the judge in the DPA hearing, implying that the judge had tested it against the evidence.

That is simply not what happens, but this misconception may cause prosecutors to become closed-minded to the problems with their cases.

Whatever the answers to the various questions raised by the disparate outcomes achieved in DPAs versus subsequent prosecutions, it is clear the DPA process allows the SFO and corporate entities to accept criminal wrongdoing on the basis of evidence which, when tested, does not stand up to scrutiny.

It should not be forgotten that in the Tesco case, the trial judge Sir John Royce ruled that the evidence was “so weak that it should not be left for a jury’s consideration” – a finding upheld by the Court of Appeal – when Tesco had already accepted in the SoF the guilt of the defendants who were subsequently acquitted.

With history repeating itself in the SGL case, it is likely all eyes will be closely monitoring the efficacy and fairness of the DPA regime in future.

Harry Travers is a partner and Alex Swan a senior associate at BCL Solicitors bcl.com

They have represented individuals in five of the nine SFO investigations connected to DPA proceedings.

Copyright & permissions