David Cameron has today announced that he is to "put rocket boosters" under the coalition's payment by results  approach to reducing reoffending. The biggest problem in penal policy is how we can reduce reoffending: the public have little time for politicians who are 'soft on crime' but they see little sense in prisons simply warehousing offenders if they come out just as or more likely to offend as when they went in. There was a danger that the departure of Ken Clarke in the reshuffle would have spelt the end of the government's focus on this critical area of systemic policy failure.
So there is much to welcome in the Prime Minister's "tough and intelligent approach" . In particular, his pledge to address the fact that short stay offenders get little by way of rehabilitation and no probation support when they leave prison has created a 'revolving door' and a cycle of reoffending.
But there was little that was actually new in Cameron's speech. There are currently four prisons pilots at Doncaster, Leeds, High Down and Peterborough. The Peterborough pilot was set up by Labour. Other areas are running community-based payment by results pilots. The coalition had already said that it wanted to see payment by results spread throughout the country by 2015.
There are two challenges that ministers must face up to if they are to make a real difference. First, rehabilitation costs in both the short to medium term. Rehabilitation requires investment in wrap around services, drug and alcohol programmes, and mental health services to provide people with the support to make a change in their lives. Yet the Ministry of Justice is facing massive cuts to its budget over the course of the current spending period. It is hard to see how a rehabilitation revolution can take off when so many local services are being cut back.
Second, we need to think about how to institutionalise a more effective and joined up approach to reducing reoffending in the long term. This is where there is a role for the new Police and Crime Commissioners. The "and crime" part of the title is important. It is plausible that at least part of the prison budget and some local prisons could be devolved to PCCs. They would then have an incentive to reduce reoffending because they would keep the savings from any fall in the local prison population. There are arguments about how much of a cash saving such 'justice reinvestment' mechanisms can yield, but the evidence from the United States where penal policy is locally administered is promising.
So the rehabilitation revolution appears to have survived the Clarke/Grayling transition. But it remains to be seen how powerful Cameron's rocket boosters actually are and whether this will produce the kind of step change we need in the offender management system.
Rick Muir is associate director at IPPR