Lies, damned lies and conviction rates

The Serious Fraud Office has been in the news this week after the publication of the HM Crown Prosecution Service Inspectorate’s recent report, Inspection of the SFO’s governance arrangements. The report is considered in detail at Legal update, HMCPSI publishes Inspection of the Serious Fraud Office governance arrangements report.

The headline of the report focuses on the perennial problem of funding, that the SFO is over-reliant on “blockbuster” funding, allocated on a case by case basis, rather than regular funding to be spent as the senior management see appropriate.

A more interesting comment emerged in the middle of the document:

“During our inspection it was proposed that the SFO calculate its conviction rate per case rather than per defendant”.

The conviction rate is currently calculated per defendant, which is the same calculation the Crown Prosecution Service (CPS) use for its cases. HMCPSI disagreed with the proposed change on the basis that it is likely to result in perverse performance outcomes and will lack consistency and evaluation with previous performance figures. They cited the example that if three cases were brought to trial with five defendants in each case and one defendant was convicted in each of the cases, the conviction rate would be calculated as 100%, despite 12 of the 15 defendants being acquitted.

The SFO conviction rates for the past six years are provided in the report:

  • 2009 – 10: 92%;
  • 2010 – 11: 82%;
  • 2011 – 12: 72%;
  • 2012 – 13: 70%;
  • 2013 – 14: 85%;
  • 2014 – 15: 78%;
  • 2015 – 16: 32%.

Commentators will naturally be drawn to the fact that the conviction rate plunges to 32% for 2015-2016. While this disappointing statistic should be recorded, and appropriate steps taken to identify where things might have gone wrong, it would be wrong to place too much reliance on a single year when the SFO’s conviction rate has been above 70% for the previous six years.

The conviction rate of a prosecution body is only one of many factors that determine its overall performance. Conviction rates are a limited measure because they are so often used as a like for like comparison in varied circumstances. In Japan, the reported conviction rate exceeds 99%. In China the reported conviction rate for all offences in 2014 was 99.9%. One could speculate on the various reasons for the high rate, but it is unlikely to be a simple matter of a more efficient and better performing prosecution agency. A  99% conviction rate of any prosecution agency in England and Wales should raise some serious concerns.

The first step on assessing the value of the conviction rate is the Code for Crown Prosecutors (see Practice note, Code for Crown Prosecutors). In short, the full Code test, most commonly applied prior to charging a suspect, requires a two stage process:

  • An evidential stage, at which  prosecutors must be satisfied that there is sufficient evidence to provide a realistic prospect of conviction. The test must be applied to each proposed charge and for each defendant.
  • A public interest stage, where the prosecutor must be satisfied that the public interest factors in favour of a prosecution outweigh those against.

The Code test is not the same as the direction given to the jury on the standard of proof, which is  beyond reasonable doubt.

Given those two positions, just as a conviction rate consistently below 50% should give cause for concern, a conviction rate regularly exceeding 85% would suggest that the Code is not being applied correctly.

The HMCSI report states that whilst the SFO is disproportionately affected by small numbers of cases in any one year, the 2015-16 results will inevitably raise concerns as to whether it has suffered a run of bad luck, the risk appetite has been set too high, or case review processes are not robust enough.

The report also acknowledges that whilst the conviction rate is a key indicator of success, deferred prosecution agreements are not calculated within the conviction rate. The work to secure a successful DPA can be just as resource-intensive as an investigation. Conviction rates should not therefore be regarded as the only measure of success for the SFO.

The question of why the SFO would want to re-define the methodology for calculating a conviction rate looks even more bizarre considering its results for 2015 – 16 would have looked weak even with the revised calculation. A per case result would result in a 100% conviction rate recorded in most years. The 2015 -16 32% conviction rate was largely the result of three cases, Swift Technical Solutions (three of three defendants acquitted), Zetland Fiduciary Group (two of two defendants acquitted) and Libor 2 (six of six defendants acquitted). During the same period four cases resulted in at least one conviction. A per case result in 2015 – 16 would therefore be 57%, a considerable drop from a normal 100% rate.

The SFO has been scarred by some questionable analysis of its conviction rate in the past. In the Review of the Serious Fraud Office published in June 2008, Jessica de Grazia, a former Assistant District Attorney in Manhattan contrasted the SFO’s 61% conviction rate over a five year period (2003 – 2007) with that of the District Attorney of New York (92%) over the same period.

The De Grazia report suggested various reasons for this difference, but failed to acknowledge the fundamental reason: differences between the standard required to meet the Code Test and the beyond reasonable doubt standard required for a jury to convict. Similarly, the De Grazia report stated that “Certainty of conviction is more important to a plea-negotiation system than the possibility of a long prison sentence”, perhaps failing to recognise the maximum sentences for business crimes in England and Wales are considerably below those that can be handed down in US courts.

However, rather than attempting to move the goalposts, the SFO’s stance should surely be to defend their rather good record over the past six years (which of course contrasts well with the 2002-2007 period) and considering what went wrong in the past year. One solution might be the adoption of one recommendation of the HMCPSI report, that “introducing some independent oversight would greatly enhance transparency and challenge”.



Practical Law David Bacon

Leave a Reply

Your email address will not be published. Required fields are marked *