It was called a social impact bond, but in many ways, as Andy said, it's more like a pay for performance model. Let me pull up my notes on that particular project.
It was an experiment that was started under the Labour government, and it's been running for a number of years. The focus of it was on supporting the St Giles Trust, which was already working in that space, and incentivizing them on a performance basis to reduce the reoffending rates.
The process started with support officers working with young offenders who were in prison for less than 12 months, and while they were still in prison, helping them to plan and to adjust to the time they'd be spending outside the prison: transitioning to housing once they left; getting engaged with the workforce; having coaching, support, and mentorship during that period. It's the younger end of the age spectrum of prisoners who are first-time offenders who have the highest chance of rehabilitation.
In terms of the costing, for that demographic typically the reoffending rate has been 75%, but for that younger demographic it had been as high as 90%. They worked with about 3,000 prisoners and were able to reduce that reoffending rate from 75% down to closer to 10%. The program officer cost around £49,000 a year. The prisoners who were in prison were costing the public purse £143,000 a year. So the 65% saving resulted in £4.7 million of savings in the course of the program.