The fallacy of falling reoffending rates
That was the description given by Antonia Romeo, the most senior civil servant at the Ministry of Justice (MoJ), to the decline in reoffending rates in England and Wales over the past ten years.
Among adults, they have fallen from 30 per cent in 2010-11 to 24 per cent in 2020-21. For those aged 10 to 17, the drop has been even more marked, down from 40.9 per cent to 31.2 per cent.
Giving evidence at the Commons Justice Committee, Romeo suggested that the reduction was because officials in her department had done “a lot of work…on reoffending and what works to get people not to reoffend any more."
She said: “This is getting them into a job and accommodation, managing the Through the Gate process, and getting them off substance misuse.”
The Justice Secretary Dominic Raab has been even more effusive.
When figures were published last October showing a two percentage point drop in reoffending levels in 12 months he tweeted: “This shows that our investment in drug rehab, training in prisons and offender employment is working and helping make our streets safer.”
Does it show that? Look closely at what reoffending rates really measure and it’s clear that the reduction is nothing to boast about - it’s simply a reflection of a wider failure to deliver justice.
That’s because the reoffending rate is a misleading term. It is not worked out by counting what proportion of offenders commit a further crime; it’s based on how many are caught and sanctioned.
Tracking the cohort
The MoJ calculates it like this.
Every three months it adds up the number of people who within that period have been: cautioned, reprimanded or issued with an official warning by police; given a non-custodial sentence at court, such as a fine or community order; released after serving a prison sentence.
This ‘cohort’ of offenders is then tracked.
If, in the following 12 months, anyone in the cohort commits an offence for which they are convicted, cautioned or given a police warning they are officially classed as a reoffender. The penalty or conviction counts only if it’s issued within the initial 12 months or a six-month period afterwards. Some offences, like breaches of court orders, don't count.
The number of reoffenders is then divided by the overall number of offenders in the cohort to produce the reoffending rate.
So, the reoffending rate is really the re-conviction or re-cautioning rate. It all depends on the offender being arrested and given a police warning or successfully prosecuted in the courts. As such, reoffending rates vary according to the effectiveness of the 43 police forces, the Crown Prosecution Service and the criminal courts.
But don’t take my word for it - even the MoJ acknowledges that the methods are imprecise and understate the reality of reoffending.
“Measuring true reoffending is difficult,” says the department’s ‘Guide to Proven Reoffending’, issued in 2017.
“Official records are taken from either the police or courts, but they will underestimate the true level of reoffending because only a proportion of crime is detected and sanctioned and not all crimes and sanctions are recorded on one central system. Other methods of measuring reoffending, such as self-report studies, are likely to also underestimate the rate,” the document says.
Detections and sanctions
To understand the extent to which reoffending rates may have been affected by the performance of police, prosecutors and the courts, just look at overall levels of crime detections and sanctions compiled by the Home Office as part of their 'outcome' figures.
The methods changed in 2014-15, so that’s the earliest comparable date. That year, 15.5 per cent of crimes recorded by police led to a suspect being charged or summonsed to appear in court. A further 4.6 per cent resulted in a formal out-of-court disposal, such as a caution.
Over the next seven years, as has been well documented, the charge and caution rate plummeted. In the 12 months to the end of September 2022 it was 5.5 per cent and 1 per cent respectively.
So, over a seven-year period there was a 3.6 percentage point reduction in caution rates and a 10 percentage point decrease in charging levels. Without a charge, of course, there can’t be a conviction, so convictions, as a proportion of all crimes, will also have fallen substantially.
You can probably see where this is heading. With such a dramatic decline in cautioning and charging rates, it’s no surprise that reoffending levels (measuring the proportion of offenders who have been cautioned or convicted again) have fallen too. In 2014-15, the MoJ calculated the overall reoffending rate to be 30 per cent. The latest figures, covering a cohort of offenders in January to March 2021 who were tracked for the following 12 months, show the rate fell to 24.3 per cent, a 5.7 percentage point reduction in reoffending.
Timeliness is also a factor. Criminal cases are taking longer to resolve because of record court backlogs which started growing in 2019. It takes around 180 days, on average, for an offence to be dealt with by magistrates compared with 150 days in 2014-15. For Crown Court cases, it’s gone up from 250 to 350 days. That means an increasing number of offenders who committed a further crime will not have been counted for the purposes of the reoffending data because they weren’t convicted within 12 months or the six month follow-up period.
If overall charge and cautioning rates had been broadly stable, then this “hugely significant” reduction in reoffending rates, as the permanent secretary put it, would indeed be hugely significant. It would suggest that external factors, such as improvements in rehabilitation, employment and accommodation support, were playing a part - but there is no evidence that they have. The evidence simply, and bleakly, points to the fact that fewer offenders are being caught and brought to justice, while delays in the criminal justice system are masking some reoffending that would previously have been included.
It’s time ministers and officials acknowledged this. They should start by re-labelling ‘reoffending’ rates to avoid confusion and misinterpretation: they are re-conviction and re-cautioning rates. Better still, they should look for an alternative way to measure the true level of reoffending through a combination of anonymous surveys of offenders, information from probation staff, arrest figures, and re-caution and re-conviction data, adjusted to take account of overall caution and charging trends.
The fall in reoffending rates is, sadly, not the ray of light the Ministry of Justice was hoping for and no one should be misled into thinking that it is.
PS: Isn't it odd that for the purposes of reoffending data, an offender's time in prison doesn't count? The official reoffending clock starts on release from jail or at the moment the court orders an offender to serve a sentence in the community. So, when comparisons are made between reoffending rates for those who've served prison sentences and people given community penalties they never take account of the time spent in custody. That doesn't seem right. One of the benefits of prison is that while locked up a person is not committing crimes in the community: surely the reoffending figures should reflect that.