The world of 2026 does not revolve around the sun; it revolves around statistics. From Pythagoras to Einstein via Euclid and Newton numbers have shaped the world we live in and know today. We, the public, unable to put our ear to the closed door whisperings of our masters have only the numbers these masters provide to us in order to make our judgements as to whether we wish them to continue with their policies for another parliamentary term. Nowhere is this filtering of information more profound than at the door in Petty France London SW1 home to the Ministry of Justice.
There are several descriptions of this filtering. It is a system where some information is readily made available, some must be extracted under Freedom of Information rules and some is kept under wraps. It could be The Empirical Boundaries of Justice; The Parameters of Judicial Transparency; The Quantifiable Outlines of Justice; The Statistical Architecture of Justice; The Informational Perimeter of Justice; The Justice Audit but I will settle for the simplicity of The Metrics of Justice.
In order to interpret the numbers available for inspection it is helpful to know how they are chosen. There are various agencies and authorities which decide which aspects of justice are important enough to measure and publish. Court and sentencing data originate largely within HMCTS and other justice agencies and are analysed and published by the Ministry of Justice's statistical and analytical teams. Professional authority over what constitutes and how official statistics are published rests with the Ministry's Head of Profession for Statistics. The wider system is overseen by the UK Statistics Authority and its Office for Statistics Regulation while the ONS independently produces the principal national crime statistics.
With millions of numbers available the decisions on which to collate and publish are made by the Head of Profession for Statistics supported by the Chief Statistician and the relevant statistical teams within the MOJ. Perhaps just as interesting to the casual observer are the statistics not published. Some that come to mind are:-
Outcomes by individual magistrates' bench;
Conviction rates by individual court;
Appeals and their outcomes by individual court;
Reasons for successful appeals;
Cases abandoned because of prosecution failures;
Defendants waiting more than specified periods;
The proportion of cases in which witnesses attend unnecessarily;
The number of cases where a legal error is corrected before an appeal;
Comparative outcomes for magistrates and District Judges.
The above list is not in any order of significance. I would think that some readers will have their own ideas of additions which they would consider to be equally or more important. With controversial changes to the jurisdiction of magistrates courts hiding behind other more pressing current matters within the MOJ there are some valuable pointers as to the quality of justice in the lower courts system. Of the defendants who appealed a magistrates' conviction in 2024 41% had their appeal allowed. Of those appealing sentence 44% had their appeal allowed. It`s interesting to note that the comparable numbers for 2010 appeals are 44% allowed against conviction and 47% against sentence allowed. Observers outside the legal fraternity might wish to come to their own conclusions about very fine disparity between those numbers considering how much water has flowed under the legal bridge in that time period.
As we go higher up the ladder of appeal the stakes for defendants, victims and the public perception become higher. In 2024–25 the Supreme Court decided 170 applications for permission to appeal granting permission in 61 (36.1%). A criminal appeal normally requires certification that there is a “point of law of general public importance” and that the point ought to be considered by the Supreme Court which is effectively the finest filter in a complex system of legal filtration. To continue the analogy a simple flow chart could be shown as follows:-
Applications → Permission refused → Permission granted → Hearing → Appeal allowed → Conviction quashed.
The subject of early release of prisoners has made headline news since the change of prime minister prompted a rocket to be sent up the arse of the movers and shakers at the MOJ. The Parole Board for the justice system can be likened to Camp IV, also known as the South Col for mountaineers. It is the last stop before the final arbiter of whether justice has been done; The Criminal Cases Review Commission (CCRC).
The statistics published by the Parole Board for 2025/26 are:-
3,140 prisoners directed for release
12,528 refused release
8,835 oral hearings
The Board says its convicted Serious Further Offences rate remains below 0.7%. Figures it supplies tell us how many prisoners the Parole Board releases and how many subsequently commit a Serious Further Offence but it leaves a gap insofar as it does not publish statistics on how many of those released prisoners subsequently commit any proven offence, how many offences they commit, how serious those offences are and whether the rate differs according to the sentence they were serving. The table below illustrates the statistics that the Parole Board does not supply.
Especially considering the current furore about early release of prisoners occasioned by the crisis in prison accommodation it must be asked whether the Parole Board`s limited public information is by accident or design.
At the end of the criminal appeals process sits an institution which is perhaps the most revealing of all when we ask whether the criminal justice system is measuring its own failures. The Criminal Cases Review Commission (CCRC) describes itself as the very last stage of the Criminal Justice System. It was established in 1997 to investigate alleged miscarriages of justice in England, Wales and Northern Ireland and where it considers there is a real possibility that a conviction would not be upheld, to refer the case back to an appeal court. It is therefore not another tier of appeal: it is an independent safety net for cases which have normally already failed at the ordinary appeal stage. The numbers are small in relation to the thousands of criminal convictions produced by the courts but they are difficult to ignore.
The CCRC's 2023/24 report recorded 1,629 applications and 25 referrals. In 2024/25 it received 1,541 applications and referred 31 cases; importantly, of the 27 CCRC-referred appeals actually heard that year, 18 were allowed and nine dismissed. In 2025/26 applications rose to a record 1,841, while referrals increased to 45, an 80% increase over the figure two years earlier.
The longer-term figures are perhaps more striking. By July 2026 the CCRC had received 35,723 applications, referred 926 cases to the appeal courts and 622 appeals had succeeded. Thus once a case has actually survived the CCRC's investigation and reached an appeal court, the outcome is very different from the initial application statistics.
This is why the CCRC should be included in any serious attempt to measure the effectiveness of criminal justice. The relevant question is not simply “How many applications does the CCRC receive?” Nor is it “What percentage does it refer?” The more revealing sequence is:
Conviction → ordinary appeal → CCRC application → CCRC referral → appeal heard → conviction or sentence overturned.
The CCRC itself makes clear that, after referral, it remains for the appeal court to decide whether the conviction is unsafe or the sentence unfair.
That final distinction matters. A justice system which records convictions, sentences and appeals but does not prominently display what happens when an independent body identifies a possible miscarriage risks measuring its activity rather than its accuracy. The CCRC is therefore not an embarrassing footnote to the system. It is arguably its final quality-control mechanism. It is in effect the Mount Everest of the appeals system.
And the question for those who decide which statistics should define the success of criminal justice is a simple one: should the number of convictions subsequently found unsafe be treated as a central measure of the system's performance, rather than as an exceptional statistic at the end of the ladder?
The figures reveal an uncomfortable truth. At every stage of the criminal justice ladder statistics are collected but the figures which most directly test whether justice has been done are often the least visible. Appeals, parole and the CCRC provide opportunities to correct error yet the system rarely presents these stages as one continuous measure of accuracy. Perhaps the fundamental question is not how efficiently the system processes cases but how often it gets them right.
