House of Commons - The role of the Probation Service

The role of the Probation Service - Justice Committee Contents

7 Practical issues in the application of payment by results to rehabilitation

245. Although there is undoubtedly scope to create a vibrant market in probation and rehabilitative services, it is clear that considerable development will be required to establish a strong competitive base that is financially stable and facilitates access to providers from a range of sectors and of a range of sizes. What it means to "create a level playing field" differs for, and within, each sector but there are also considerable cross-sectorial commonalities in what is required, no matter what level new commissioning arrangements are introduced at. Creating a market based on payment by results is also time-consuming.

246. Several of our witnesses commented on the complexity of developing a commissioning model based on payment by results. Helen Edwards, Director-General for Criminal Justice at the Ministry of Justice, gave us her views: "Although there is a good deal of agreement, or certainly we feel there is, around the payment by results approach and the concept, it is complicated and getting it right will be quite difficult."[390] Time is needed to develop market and complexity of devising a system of payment by results which Peter Neden of G4S believed would need to evolve: "the answer isn't obvious and immediate. It might be a simple concept, but delivering it in reality is quite complex. The second point I would make is that it is not all or nothing. We haven't gone from a world where people are just paid money for trying to do things to a world where we are only going to be paid for outcomes. There is a transition period that allows us as providers to become more confident about the payment-by-results system and to build the evidence base that is in the interests of both our clients and us as providers to be able to make those steps."[391]

247. John Thornhill described to us a series of concerns that sentencers would like to see answered as the Government designs the new commissioning model:

We raised the concerns [with the MoJ] about cherry-picking. We raised the concerns about how far this will be driven by profit rather than by effective delivery of the sentence. We raised the concerns about what would be the consistent approach across the country. For instance, what about the demographic issue? It will be attractive to bid for a contract in, say, somewhere like a large conurbation where there may be a wide range of clients to deal with. But in the country areas, again, what would be the difficulties there? How will they deliver across the whole of a large area? [...] We [subsequently] put a paper to the Ministry of Justice where we expressed the view that there needed to be very clear quality assurance and quality control built into the contracts, with regular appraisals, inspections, and the ability to withdraw from a contract—a whole range of commercial statements that you would have in a contract.[392]

We discussed many of these matters with our witnesses and here we draw conclusions about some of the key issues that the Government will need to resolve as it creates a payment-by-results based commissioning model.

The creation of a stable market

Immaturity of the current market

248. UNISON described the existing market as "immature".[393] We were also told that NOMS has failed to establish an effective market in probation services and there continue to be shortcomings with existing national contracts with external providers.[394] Jane Coyle of Blue Bay Support services was concerned that in her experience of engaging with the probation service to negotiate to provide more cost-effective pre-sentence reports and offending behaviour programmes, there is a lack of will to diversify the market.[395] Roger Hill had similar concerns:

What matters is a construct that is based on a belief that a range of providers is a good thing for the system. I think the probation service are quite frightened of competition and, although they may describe a view that is similar to what I have just said, they don't really mean it. So long as we ask the system to reform itself, in my view, that reform is unlikely to happen.[396]

However, he further explained why he believed the market had not developed thus far:

The reason that there isn't a market and there isn't active competition is because the commissioners are not putting work to the market. It is providers and, as providers, we can influence that and we do it on a weekly, sometimes daily, basis. We are very keen to compete amongst ourselves and with the public sector. But until the commissioning decision is made, rather like the example I gave of community payback—and they are difficult decisions to make; I know they are because I have been on that side of the fence—there is any number of views as to why you shouldn't do it or, if you have to do it, why you shouldn't do it now. Those decisions have to be the first ones. Once you start to make them and become more confident, and NOMS becomes more confident as a commissioner, and I think it will under this new construct, then that market will develop. That is how markets do develop.[397]

249. Other witnesses suggested that this may have as much to do with a lack of central drive to open up the market as individual probation areas' reluctance to do so.[398] Martin Narey explained to us that while he was CEO of NOMS there was a shift to less central enthusiasm for contestability.[399] On the other hand, the impact of competitive forces and the threat of transfer can, in themselves, spur improvements within existing service providers.[400] We have also heard that the financial situation that trusts and their local strategic partners now find themselves in has already encouraged greater external commissioning, both by individual trusts, as they have reviewed their core business, and by local partnerships, as they have been motivated to pool resources.

250. There are differences across the various sectors that must also be considered. For example, Clinks has suggested that there was a need to address cultural issues between the public and voluntary sectors.[401] Private sector procurement specialists with experience of working with NOMS, Excalibur, expressed concerns that the differences between dealing with prison-based and community-based offenders are not fully appreciated, either within NOMS and the MoJ, or by potential new providers.[402] As we noted above, Napo has raised similar concerns about the cultures of prisons and probation, which would have implications for the design of a new commissioning model.

251. There is likely to be considerable variation across England and Wales in the work that will be required to establish a stable market. In some areas there will be a local infrastructure of existing organisations but in others this will need to be developed. For example, some trusts noted that voluntary sector providers would need to build up their capacity to take a greater role.[403] In addition, there is a strong likelihood that the economic situation will have a considerable impact in diminishing the range of providers whilst there is an increase in demand. We heard that this is a particularly "big issue" for the voluntary sector. For example 60% of Clinks members were using free reserves to finance frontline delivery.[404]

252. The MoJ must not underestimate the work required to create a stable market and it must take into account the existing cultural differences across the different sectors and the complex nature of probation work. There is also clearly a possibility of early commissioning failures and, in order to maintain public confidence in the system, the MoJ must have clear contingency plans for dealing with any such problems which may arise. Innovation involves risks and those risks need to be managed.

TRANSPARENCY AND AN EFFECTIVE DIALOGUE

253. According to the CBI, a prerequisite to the creation of effective markets is for commissioners to establish and maintain "open communication" with potential providers, for example, to ensure that contracts are acceptable to the commissioner and the market.[405] Roger Hill explained why he felt communication and consultation were important: "[Sodexo] would like to see the continuing competition of public sector prisons and we would like to be part of constructing how payment by results, locality based commissioning, place based budgets and all of that develops, because we take the view that there isn't a straightforward simple answer as to how to do that. It requires a lot of thought and a lot of discussion and we would like to be part of it."[406] Catch22 advocated earlier involvement of the voluntary sector in the development of new commissioning practices.[407] There does appear to have been communication between NOMS and providers from the private sector on the new commissioning strategy.[408]

254. Conversely, the Probation Association expressed disappointment that NOMS had not consulted them about either the tenders for community payback or the forthcoming commissioning strategy.[409] Sebert Cox, Chair of the Probation Association, explained the implications of NOMS' lack of consultation with probation trusts with respect to the community payback tendering process: "The concern here is that we have been thrust into this particular competition without any consultation whatsoever. If we had had that, we might have been able to advise on the best route to take to end up in a place where we will save money and possibly keep the localism agenda intact, because lots of work has gone into ensuring that services on the ground have been properly thought through with partners and so on".[410]

255. There is a need for open communication with providers from all sectors to inform the commissioning strategy as well as in any future competitive processes to ensure that the best balance can be struck between efficiency and localism. We are concerned that this does not appear to have been the case to date.

IMPLICATIONS FOR TRAINING

256. Unlike the work of prison officers, which our predecessor Committee concluded was best informed by "life experience" rather than extensive formal qualifications, a thorough supervision and training process for probation work is necessary to equip professionals with the skills and knowledge that they need to exercise expert judgment and discretion in managing offenders safely in the community and to provide a basis for consistent performance.[411]However, the qualifying framework is not currently easily accessible to candidates outside the statutory sector, and there are no equivalent qualifications available to practitioners in private or voluntary organisations.[412] The infrastructure for training would also need to be re-considered, as few regional training consortia now exist and regional training arrangements rely on support by chief executives of trusts.[413] There is similarly a need for consistent post-qualification arrangements.

257. The Probation Chiefs' Association stated that trusts already have the business skills to commission offender services, although some trusts suggested that additional training was required.[414] When we asked Robin Wilkinson, Director of NOMS Human Resources, what training he envisaged would be required to equip trusts under any new commissioning arrangements he responded: "there has already been quite an investment in thinking through what commissioning means, and in training and developing key staff within probation trusts around those skills, but it is clearly going to be an on-going part of our world."[415]

258. The success of any new commissioning model in protecting the public will be predicated on the existence of strong safeguards to monitor standards of professional expertise. Staff undertaking offender management work on behalf of other sectors will require the same high-quality qualifying training as probation professionals working for trusts. In order to foster some consistency in the specialist skills required to work with particular types of offender, and a high-quality service provided, we would like to see the MoJ working with Skills for Justice to create a framework for both accredited qualifying and post-qualifying training that is accessible to all providers.

The complexity of designing a payment by results commissioning model

259. Helen Edwards described what she foresaw as the role of the Ministry of Justice under any new commissioning arrangements:

We will still have responsibility overall for standards. In the way we put contracts together, we will want to be clear about the parameters and who is responsible for what. We have yet to work out the detail of that, but I think, overall, the broad approach is for more discretion to be exercised by those who provide services, and with that goes increased responsibility for what is done.[416]

260. We heard that working through this detail was likely to be incredibly complex. Witnesses have raised questions about the implications for the application of payment by results to criminal justice of: the risks of failure in terms of public protection; the validity of indicators of reduced re-offending; the costs to providers of monitoring and the national costs of commissioning; ownership and commercial confidentiality of information; and the quality of the evidence base on which to base commissioning decisions.

THE FLEXIBILITY OF CONTRACTS

261. There will be a need to strike a balance between over-prescriptive contracts which may inhibit innovation and including sufficient safeguards to promote quality services that are effective at both reducing re-offending and protecting the public, particularly in the context of fewer national standards.[417]We heard that in other fields a so called "black box" approach had been adopted whereby "the commissioning department defines very clearly the high-level outcomes it wants to achieve...but they leave it up to a provider to then determine how they are going to find a solution to that"; the requirements in contracts generally have been limited to legal requirements to deliver a minimum service.[418] On the other hand, while part of the potential value of payment by results is in the transfer of risk and responsibility to the provider, there is a danger that public protection will be compromised if contracts are not properly designed. Mr Eccles of Social Finance identified potential tensions arising in applying such a model to some areas of probation work as public safety concerns may increase the extent to which service provision is specified by the MoJ.[419] However, in practice it may not be easy to draw such distinctions, as the risk presented by an individual offender does not remain static throughout the duration of the order.[420] It is possible, however, to build a set of principles into the contracting process, including for example, safeguards around working with vulnerable individuals, including victims, engagement with sentencers and treating people with dignity and respect.[421]

262. Contracts also need to be flexible enough to ensure that providers address offending-related needs as and when they arise rather than purely on the basis of the sentence of the court and any assessment of needs at the time of the order. Probation trusts frequently go above what is required in their service level agreements with NOMS, by providing a greater volume of interventions than they are contracted to and by engaging in partnerships to work with offenders who are not their statutory responsibility.

263. The Ministry of Justice will need to ensure that contractual specifications include adequate safeguards for public protection. The current work of probation services is not just about meeting the obligations in their contract; its work is about serving the demands of the courts, which cannot be easily predicted, and being sufficiently flexible to meet the offending-related needs of individuals as and when they arise. The Government must clearly specify in its new commissioning strategy how it intends to ensure that contracting arrangements with providers from other sectors will accommodate these needs and respond to unexpected changes.

Diversity

264. The progress that probation trusts have made towards more equal treatment in the criminal justice system could be undermined if contractual safeguards are not put in place. Probation trust performance analysis incorporates diversity criteria to enable improvement. Avon and Somerset Trust considered that it will be important for any providers of probation services to at least meet the current attention paid to diversity and equalities.[422] Nevertheless, the green paper does not comment on the implications for addressing diversity of contracting out services and the potential loss of the public sector equality duty; neither is it clear whose responsibility it would be to monitor practice in this regard.

265. From April 2011 the public sector has had a statutory duty to take positive action to eliminate gender discrimination and promote equality under the Equality Act. We would welcome a clear statement in the commissioning strategy about how the equality duty will apply to providers outside the public sector and how the Government intends to ensure that probation and rehabilitative practice is fair and inclusive, particularly in the context of increased provider and professional discretion.

DESIGNING APPROPRIATE PAYMENT MECHANISMS

266. The Government's proposed model for the application of payment by results to community sentences is to make an initial payment to providers for the delivery of community orders and a subsequent one for reducing reconviction, yet it does not intend to provide additional resources up front to finance local incentive schemes. We had a great deal of discussion with our witnesses about their views on the key components of effective payment mechanisms, including the appropriate results, to ensure that they are attractive to potential providers; are financially viable; they facilitate the participation of a range of cross-sector providers; and provide the right incentives to ensure that the aims of the criminal justice system are met.

Access to working capital

267. Access to working capital was raised as an issue which could undermine the ability of organisations across all sectors, particularly small and medium sized organisations, including probation trusts, to participate under both the models described above. For example, trusts themselves are not in a position to bear the financial risk of investing money in services to facilitate payment by results schemes, either in terms of having the sufficient funding to invest upfront in new services or to sustain themselves in the event that they do not get paid, a problem exacerbated by the fact that they are not allowed to hold reserves.[423] For example, Clinks identified that "risk appetite will be very low at the local level where [VCS] organisations have little or no reserves" and Catch 22 argued that access to capital from mainstream financial services which, it stated, are "much more geared towards catering to the private sector".[424] Cash flow, and the level of financial risk, is also an issue for both small and large private sector organisations, for the latter particularly where it is difficult to make informed investment decisions as a result of a lack of evidence.[425]

268. While witnesses welcomed the Government's proposed model of payment for community sentences, described by Christine Lawrie as a "core plus" payment system, we heard that to make payment by results viable for any organisation it will be important to get the ratios right between upfront "guaranteed" payment to provide working capital and the element attached to "results".[426]

269. Social finance funding for the Peterborough project has come from trust funds and the Big Lottery Fund.[427] However, Andrew Neilson of the Howard League, Clive Martin of Clinks and A4E were sceptical that there would be sufficient social investors interested in criminal justice in the short-term.[428] Andrew Neilson warned that, in his view, payment by results pilots, and the six year social impact bond project in Peterborough in particular, will not be able to prove itself in sufficient time to attract new money into the system; payment by results is not going to reduce the numbers in the criminal justice system any time soon.[429] There may also be difficulties in designing financially viable schemes as the minimum size for a social impact bond model is dependent on what is statistically significant in terms of reducing reconviction for the target population as well as the size of the cashable benefit of that reduction. Rob Smith of YSS explained that West Mercia had been trying to develop a similar project for two years but "the scale you have to get to in order to make the cashable savings for the Treasury appears so large and is very difficult to overcome. Agreeing a set of metrics to a more rigorous standard that had probably been initially worked out for the Peterborough model is making it impossible to develop anything at present."[430]

270. Crispin Blunt was more hopeful about the role that private investors could play in bringing new resources into the system:

I would sincerely hope that it will not just be Social Finance coming forward to finance schemes like Peterborough, and social philanthropists and charitable trusts investing in Social Finance to deliver Peterborough. I sincerely hope that with our payment-by-results schemes we will be able to convince the market that this is a proper investment vehicle, because they should be sufficiently confident that they are going to add to our capacity to drive down reoffending and so improve the people in the care of the state, either in custody or in the community, that we are going to be able to take the savings that they have invested in creating the capacity to do it. That has to be the objective.[431]

271. Our predecessor Committee considered the relative merits of performance incentives versus financial incentives for driving reductions in re-offending to the extent required to prevent prison population growth. It concluded that additional financial incentives had the potential to accelerate progress at local level; evidence suggested that this would require some pump-priming.[432] There is considerable scope for partnership approaches to provide viable alternatives to short-custodial sentences but these can require high initial investment, and Hertfordshire Probation Trust questioned whether the local incentives model would facilitate their development.[433]

272. When we questioned the MoJ about the viability of the local incentives model when no additional resources would be provided, it recognised the difficulties of organisations getting working capital to finance start-up costs but was clear that it did not expect to have to provide the resources to underpin service provision.[434] The Youth Justice Board has taken a different approach—more akin to our predecessor Committee's recommendations—and created a small reinvestment fund.[435]

273. We see that there is considerable potential in social finance, but it will take time to develop. The Ministry of Justice should learn from the experience of other Departments including the Department for Work and Pensions and the Department of Health.

CHOOSING APPROPRIATE OUTCOME MEASURES

274. The complexities of devising appropriate outcome measures on which to base payments in criminal justice must not be underestimated; the relationship between intervention and outcome is less straightforward than in other fields.[436] It is the Government's intention that the main 'result' that would attract payment is reduced reconviction. However, reduced re-offending is by no means the only purpose of a sentence: courts must also consider appropriate punishment, public protection, general deterrence and reparation to the victim. This raises the question of whether it is appropriate to assess an intervention in terms of a single purpose when it may have been intended to achieve others.

275. There are also questions about the reliability of the various indicators of reconviction and about which would be most appropriate for payment by results. Reconviction is used as a proxy measure for reoffending because it is the best indicator available, although there are obvious limitations in its use, not least that recorded crime accounts for between only a quarter and a tenth of total crime. Levels of reconviction also vary by offence and by police area, so that variations in rates may reflect variations in policing performance; and they are dependent on the efficiency of the courts.[437]

276. As a result of attempts to improve the sophistication of reconviction as an indicator of re-offending, a number of measures have been used in recent years. Examples include a measure of the volume and severity of offences committed, and a local re-offending snapshot measure, which takes a cross-sectional approach to examine the reconviction rates of the caseload of each trust on a given day.[438] Probation trusts currently use the snapshot measure coupled with a series of proxy measures, including successful completion of orders or licences and entering employment or stable accommodation.[439] The Ministry of Justice has identified three indicators which could be used as an outcome measure of reduced re-offending: 1) All proven re-offending, including reconvictions, cautions and penalty notices for disorder; 2) All proven re-convictions; or 3) Reconvictions leading to a serious disposal i.e. community order or custodial sentence.[440]

277. Witnesses explained that measuring re-offending is very complex and that, as a result, there are limitations with each measure.[441] The most robust or accurate measures cannot be provided until a year or two years after sentence; this time-lag would have important implications for cash flow in a system of payment by results.[442] Martin Narey explained why he favoured frequency measures: "What we need to know when we are looking at what we do with offenders is whether they offend less, both in terms of quantity and gravity of offending. If someone leaves a prison after a 10-year sentence for an armed robbery and is convicted within the next two years for a petty theft, that is seen as a failure, when actually that is probably a great success."[443] Another potential issue is that re-offending is currently measured against a national baseline which would fail to account for existing local variations in performance, yet the existing local indicator used by trusts—the snapshot measure described above—is not comparable to the national baseline.[444] Nationally, we were told that the snapshot measure is indicating a "statistically neutral" impact, i.e. that probation trust practices are having neither a positive nor negative effect on re-offending.[445] South Yorkshire is reportedly the only trust where there has been a statistically significant reduction in re-offending levels since the introduction of the snapshot measure: over 90% offenders do not re-offend.[446] The snapshot measure is currently being reviewed by the MoJ. Hazel Kemshall advocated examining the impact of local probation practice on re-offending by examining actual reconviction against what it is predicted those offenders would do based on risk assessment measures in OASys.[447]

278. The choice of indicator of reconviction is important so as to ensure that providers are adequately incentivised to target the right offenders. The Ministry of Justice also noted that there may be a case for rewarding providers differently for rehabilitating different types of offender, for example, female offenders, offenders from minority ethnic groups and young adults.[448] If this were to be the case it may be necessary to use different indicators of re-offending for different types of offence and over different time periods.[449] Clive Martin argued that measures would need to vary by type of offence, for example, sexual offences, would require a binary measure, but for a persistent drug user, for example, it would be unrealistic and may require intermediate measures e.g. completion of a drug treatment programme.[450]

279. We welcome the Ministry of Justice's review of the local snapshot measure of re-offending. Once a new, more robust, measure is devised, it would be prudent for the Ministry of Justice to commission research to examine practice in the best and least performing trusts so as to strengthen the evidence base.

Interim/proxy measures

280. As a result of the limitations of reconviction measures, several witnesses called for the use of interim or proxy measures for those factors proven to have a critical role in reducing re-offending as a sounder basis for payment by results.[451]For example, Hertfordshire Probation Trust proposed that payment could be determined by achieving specific measurable results such as commencement and completion of treatment or intervention or commencements and completion of qualification with an overall supplementary payment linked to re-offending.[452] These measures are useful, particularly in the absence of a strong evidence base, as there is a relationship between an offender's compliance with their sentence and behavioural change.[453] However, the factors known to be associated with offending are not separate but inter-related—for example, being without employment makes it hard to find settled accommodation and vice versa.[454] Martin Narey emphasised this point, telling us that

Despite the flaws in the research, what the research tells us over and over again is that, if you get a prisoner, or someone on a community penalty, somewhere to live and into employment or training, they will offend much less, by about a half. There is lots of analysis to prove that. A valid measure of success for community penalties or for prisoners would be having somewhere to live and in a job after release. That would give you a dependable indication of the likelihood of someone not reoffending.[455]

GRADUATED AND DIFFERENTIATED PAYMENTS

281. There is also the question of how best to incentivise providers to avoid the risks of "creaming and parking" of potential clients—i.e. focusing their services on those who would be most likely to achieve the financial outcome—which has been a problem with models in other fields, notably under DWP programmes.[456] This is particularly important in criminal justice as "cherry picking" offenders would represent a potential threat to public safety.

282. Councillor Khan summarised the potential hazards in using inappropriate measures to determine payment: "Whatever size of contract is given, if there is the same tariff for all the individuals, I just think if I had this limited resource to work with that client group I am going to try and get the most bang for my buck on that, which is why differential payment tariffs are really important on that."[457] It is thus important that the "result" upon which payment is based is carefully calibrated to ensure that providers cannot choose to work with the "easiest offenders".

283. Other witnesses similarly suggested that differentiated payment models should be devised to incentivise providers to work with those offenders who may be harder to rehabilitate. For example, under the DWP Work programme initial payments, or "attachment fees", are made on the basis of eight categories which relate to the type of customer, including for example, lone parent, ex-offender, young person, and the duration of a person's unemployment; there are subsequent weighted payments on a pure outcome basis that are differentiated again by a customer cohort basis.[458] The Department of Health has designed a complex tariff for mental health care under which payments determined using a set of 21 'care clusters' that together form 'currencies' or units for contracting; each cluster defines a group of service users who are relatively similar in their care needs and therefore in their resource requirements. It is worth noting that it took considerable time to devise these metrics.

284. However, we heard that it can be difficult to define "who is hard to help or how they are hard to help, and how much to pay extra for those who are hard to help".[459] Hertfordshire Probation Trust suggested that a supplementary payment linked to re-offending levels could be used to incentivise partnership co-operation, whereby all agencies are working to an overall shared outcome. It also proposed a banded system of differential payments to encourage work with range of offenders e.g. first-time, intermediate, prolific and high risk of harm.[460] Another option would be to align the bands with the four tiers of the offender management model, which rank offenders according to their risks of reconviction. For example, Councillor Khan explained:

One possible solution is that, for those clients who have the highest risk of reoffending, their payment by result tariff is larger than for those who are at a low risk of reoffending. Therefore, within the system, regardless of which client group you are working with, the input you put into that particular client and the output that you receive is commensurate. That kind of system is going to be more difficult and bureaucratic to manage, but I believe in the longer term it will deliver the best results.[461]

285. There are also examples of graduated payment for achieving a mixture of activities and outcomes. Chris Wright drew our attention to Catch 22's INSPIRE youth resettlement programme under which it receives payment on the basis of: 10% on contract signing; 15% on starts; 25% on positive activities or skills development; 20%* entry to employment, 30%* on sustainable job outcomes (*or sustained education or training if under 16).[462] The Peterborough social impact bond pilot is tasked with reducing the frequency of offending for a whole cohort of ex-prisoners which provides a natural incentive to work with those most likely to re-offend.[463]

286. As outcome-based payment models in other sectors have evolved, differentiated payments have been introduced to suppress the propensity of providers to focus on those who are easiest to help. We urge the Ministry of Justice to ensure that any model of payment by results builds in incentives for providers to work with all offenders from the outset, and that interventions are especially targeted at the riskiest offenders. We favour the development of a graduated payment system which includes some upfront funding, with additional payments based on proxy measures and then ultimately on reduced re-offending.

Developing a business case

287. The Government believes there is a strong economic and social case for investing in rehabilitation and scope to increase the use of rehabilitative requirements in community sentences, yet at the time that the green paper was published, it had yet to develop the business cases for the individual rehabilitative proposals.[464] In order to develop a business case an analysis of the costs of existing services is required, which can be built on—once evidence on outcomes, and their costs, is available—to generate a cost-benefit analysis.[465]

288. Work is underway to improve NOMS' information on costs to provide a framework for future commissioning. By April 2012, the Specification, Benchmarking and Costs (SBC) programme will have developed: a directory of over 70 services that NOMS funds for delivery to offenders, defendants, victims and courts so that costs and performance can be compared; service specifications with clear outcomes and outputs, so that any provider can deliver services under service level agreements or contracts; operating models to help determine what services should cost and show how the service could be delivered effectively and efficiently; and direct service costs to allow NOMS to understand the impact of commissioning decisions. However, Ms Byrne stated that, despite having initial costings, it has taken at least 15 years for the Department for Work and Pensions to be "reasonably confident" that they have a "fairly accurate" calibration of cost.[466] Martin Narey admitted that in the early stages of prisons competition there was not a very sophisticated understanding of public sector costs.[467]

289. Crispin Blunt acknowledged that building payment by results commissioning model would be a learning process:

In the course of all of this, people will make losses and some people will make supernormal profits, because people will make judgments both in writing the contracts in the public sector and in taking up the contracts in the private sector which they will get right and get wrong. We are just going to have to learn as the market develops to make sure we get closer and closer to what should be a proper rate of return because it does not suit anybody if a company makes a disastrous decision, loses its shirt and then finds it is in difficulty delivering its contract. That does not help us any more than it does if it then becomes a national scandal that this company appears to be fleecing the taxpayer by having taken them to the cleaners in writing the contract.[468]

290. Witnesses identified some problems with basing a new commissioning framework on the costs determined under the SBC programme. Some trusts, including Wiltshire and Staffordshire, noted difficulties in achieving increased innovation within the context of the highly prescribed service specifications which they are currently subject to.[469] Furthermore, Durham Tees Valley Probation Trust argued that the SBC programme has taken little note of the effectiveness of services delivered.[470] Cambridgeshire and Peterborough Probation Trust therefore expressed a need for a realistic specification of services to be provided.[471] The SBC costs were also based on service specifications that were based on the old national standards. Under the Pathways to Work scheme the contracting price was determined through a more iterative competitive tendering process where tenders are submitted and defined by those who are tendering, and then assessed as part of the competitive process.[472]

291. The introduction of payment by results into criminal justice is intended to be cost-neutral.[473] Nevertheless, while there is the potential for cost savings to stem from driving out existing inefficiencies, and from 'upstream' savings by reducing reoffending, there is also the prospect that costs will increase. For example, UNISON believed that such a system would require additional bureaucracy for the oversight of commissioning arrangements, contract management and quality assurance which would divert resources from the frontline.[474] Indeed the application of payment by results in health and employment resulted in increased costs. For example, administrative costs were estimated to have increased by around £100k-£180k in hospital trusts and from £90k to £190k in primary care trusts (PCTs). Most of the additional expenditure was due to: recruitment of additional staff; higher costs of data collection; higher monitoring costs; and higher enforcement costs. While there are lower costs in negotiating prices and volumes, this is offset by difficulties in managing activity levels.[475]

292. There would undoubtedly be some cost implications for the application of payment by results to criminal justice. For example, Tessa Webb, Chief Executive of Hertfordshire Probation Trust, stated that the probation service does not currently have access to the police national computer which provides data on re-offending.[476] Sonia Crozier, Chief Executive of Sussex and Surrey Probation Trust has commissioned research to create a costing framework to assess the impact of local delivery models which would be required to facilitate the implementation of payment by results.[477] We were told that other than for the evaluations of the pilots, the MoJ has not budgeted for an increase in administrative costs; indeed the administrative costs of the Department are being reduced by 30% over the spending review period.[478]

293. Nevertheless, it is also important for the Government to reflect on the potential benefits of greater investment in generating better outcomes. Mr Eccles raised concerns that consideration of the merits of any new commissioning model on cost alone could overlook the need for contracting processes that deliver higher quality services and better outcomes; it should focus on value for money.[479]

294. In any new commissioning model there is a need for a balance to be struck in ensuring that the administrative costs of any new system do not outweigh the potential benefits. The Government's pilots should be designed to enable them to create a business case for a viable model of justice reinvestment to be implemented over the next spending review period.

THE QUALITY OF THE EVIDENCE BASE

295. For any provider, judgments about the type of services to commission need to be based on the best available evidence if they are to have optimum impact on the reduction of re-offending, and hence on being paid. Although there has been a major drive to implement effective practice in terms of the delivery of probation interventions, the evidence of programmes having a demonstrable impact on reconviction is patchy, and the emerging importance of desistance theory, which offers an alternative understanding of how best to stop offending, is largely untapped, and therefore untested.[480] Mr Eccles summarised the situation: "while there is a lot of data available to predict whether people will reoffend or not, there is less data available on interventions that will work to stop them."[481] Mr Martin identified a risk that commissioning decisions may be "arbitrary and based on almost historical practice.[482] There is also a risk that this will undermine the ability of the MoJ to bring new money into the system.[483]

296. Professors Hedderman and Pease explained that there are limitations in the methodology that enables comparisons of the effectiveness of custodial and community sentences, as reconviction is counted from the day of release for imprisonment and from the date of sentence for community sentences, yet they disagreed about whether a uniform measure should be based on the beginning or end of the sentence.[484] They also had different interpretations of the evidence that was available to compare these sentences. Professor Hedderman cited Ministry of Justice evidence that "seems to show" that the impact on reconviction of probation and suspended sentence supervision is greater than a short prison sentence. On the other hand, Professor Pease's interpretation of other available evidence from the MoJ was that it "appear[ed]" that community sentences make no difference to reconviction as "one can predict the probability of reconviction on the day of sentence as well as one can at the end of sentence"; he also cited similar international evidence from the limited number of randomised controlled trials comparing short prison sentences and community payback that have been conducted in this field.[485] Professor Hedderman was sceptical of the value of randomised controlled trials in evaluating the effectiveness of sentencing because sentencing is not random, i.e. the particular circumstances of the individual form an important part of the sentencing process.[486] While there are some cost-effectiveness evaluations, notably by Matrix, that indicate that community supervision is more cost-beneficial than prison, the methodological issues described above have implications for the ability of providers to understand the relative cost-effectiveness of particular interventions.[487] The Ministry of Justice needs to develop a measure that enables the effectiveness of prison and community sentences to be compared more robustly.

Potential to strengthen the evidence base

297. The Secretary of State believed that payment by results provides an opportunity to strengthen the evidence base: "If people start making claims and we contract for it, one minor side effect might be that we get improved information about what actually is happening in terms of re-offending and different types of treatment."[488] However, Mr Eccles of Social Finance believed that while payment by results offers the potential quickly to build a large set of evidence, realising the benefits of the data is dependent on ensuring that providers are willing to, or contracted to, make the information generated public.[489] Mr Ledger, General Secretary of Napo, explained that "[the probation service] is very much based on altruism and the belief in co-operation, and the sharing of ideas and views", thus where academics or probation practitioners have developed work it has been shared: "it has not been something to sell or to keep to yourself in order to compete over it".[490] According to Helen Edwards, the Ministry of Justice is currently adopting a collaborative approach to information sharing within pilot projects but acknowledged that "once you go into a competitive process it gets more difficult."[491] The Ministry of Justice is funding the evaluations of the payment by results pilots and will in future have responsibility for disseminating good practice. It is not clear whether national or local commissioners will be expected to evaluate the outcomes of payment by results contracts once they are rolled out.

298. It could be argued that the idea of payment by results is predicated on a notion of cause and effect that does not fit well with what is known about the nature of desistance i.e. stopping offending. Interventions do not 'cause' people not to offend. What is known about the processes of change is that desistance is not a linear path, it is characterised by lapse and relapse and a results framework could struggle to comprehend that. Tessa Webb explained: "you have to weigh up outcomes in a variety of ways. That is one of the challenges that the probation service really struggles with […] we see achievement with offenders, but it is two steps forward and one step back […] it is about how you measure progress. Desistance is the ultimate goal."[492] Nevertheless, Professor Pease believed that incentivising probation services and other providers potentially offered a means of strengthening the evidence base on the interaction between the qualities of staff and behavioural change leading to reduced reconviction.[493]

299. Nevertheless, while it is difficult to establish firm causal links, several trusts gave us examples of reductions in re-offending, and resulting cost savings to the public purse, as well as to individuals and households, that seemed to be strongly associated with changes in practice, including prolific and priority offender schemes, intensive offender management schemes and transition to adulthood programmes.[494]It is harder to translate these into cashable savings from the prison budget, although this had been achieved by the Youth Justice Board.[495]

300. It is important that data and information sharing is not inhibited by the rules governing commissioning and competition.

ATTRIBUTING RESULTS

301. The MoJ noted that "[a] particular challenge in developing [payment by results] is ensuring that the results delivered are attributable to the work of the provider and do not generate any adverse outcomes."[496] Witnesses agreed, particularly in the context of a range of payment by results programmes being introduced across Government.[497] Professor Chalkley explained how this challenge had manifested itself in health: "There has been big literature in health care about the practicality of paying according to genuine outcome, as in whether the treatment is successful. That literature emphasises the difficulties that would be involved in doing that: the difficulties in establishing whether it is the actions of the hospital or the actions of health carers that have given rise to the outcome rather than just chance and good or bad luck."[498] However, the Secretary of State told us that "You have to put up with the fact that if you pay people because someone does not re-offend, in some cases they will be just lucky because nothing they do has contributed but the man will not offend again."

302. Mr Blunt displayed a similar lack of concern about the potential for Government to be making duplicated payments for work with an individual offender across several different payment by results schemes:

If there is overlap—someone takes a drug-addicted prisoner and turns him into a tax-paying employee, and they get the bonus for getting someone into work; they get paid for getting a drug addict clinically independent of drugs; and they get paid for stopping an offender reoffending, so they manage to score under three different schemes—I suspect that is probably a perfectly satisfactory state of affairs [...] It is our job to make sure that one of the schemes isn't paying for all three of them up front and they are getting paid double underneath.[499]

303. Effective application of payment by results measures must be sensitive to what is and what is not within the control of the provider, otherwise there is a risk of: perverse incentives to manipulate results to meet a payment measure, for example, not to arrest or not to breach; or inaccurate conclusions, if the positive results are not related to the intervention.[500] Hertfordshire Probation Trust explained that there was potential for the police not to arrest in order to meet a payment by results measure, for example. Inaccurate conclusions could be drawn about effectiveness if positive results are not related to the intervention.[501] Perverse incentives might also hinder effective approaches. For example, Hertfordshire's 'Choices and Consequences' programme requires offenders to take responsibility by admitting to the full extent of their crimes, which would adversely affect the reoffending rate.[502]

304. The Department needs to address the risk that providers may receive multiple payments under a range of payment by results programmes, for example, for employment, drugs misuse and reducing re-offending.

390 Q 258 Back

391 Q 617 Back

392 Q 575 Back

393 Ev 184 Back

394 See e.g. Q 423 [Mr Wright]; Ev w134 Back

395 Q 596; see also Ev 223 Back

396 Q 604 Back

397 Q 611 Back

398 See e.g. Ev 196 Back

399 Q 463 [Mr Narey] Back

400 Ev 200, citing Julian LeGrand; see also Q 478 [Mr Narey] Back

401 Clinks, Response to Breaking the Cycle consultation, March 2011 Back