Abstract
In recent years the UK government has emphasized evidence-based policy, as a part of which the Payment by Results (PbR) approach has increasingly been promoted. Payment by Results allows the government to pay a provider of services on the basis of specified outcomes achieved rather than the inputs or outputs delivered. Linked to PbR is the innovative source of funding social interventions know as Social Impact Bonds (SIBs). We discuss the potential benefits of PbR and survey its use across the UK public sector. Then, concentrating in particular on the Criminal Justice System (CJS) we outline three methodological challenges to the implementation of PbR: gaming, measuring outcomes and identifying fiscal benefits. We then identify a number of challenges to the implementation of PbR. We conclude that PbR has a place in commissioning services, but that its role in the CJS is likely to be limited and unlikely to succeed as a socially efficient means of attracting new sources of funding. We finish by considering two alternatives to PbR: justice reinvestment and personalization.
Introduction
Payment by Results (PbR) ‘will link payment to the outcomes achieved, rather than the inputs, outputs or processes of a service’ (Cabinet Office, 2011: 9). The idea of paying for specified outcomes, which will, it is hoped, reduce government costs, is particularly attractive at the moment because of the current state of public sector funding in the UK.
Because payment arises after outcomes are known, which might involve substantial time delays, a key challenge to the delivery of PbR models of service delivery is the challenge of raising working capital (Mulgan et al., 2010). A related policy innovation, Social Impact Bonds (SIBs) will allow the government, in theory, to privatize the up-front costs of social innovations and the associated risks, thus reducing taxpayer expenditure in the short-term and eliminating the risk of government money being spent on interventions which do not deliver the desired outcomes. A social investor might be, for example, an individual philanthropist or a charitable trust. Mulgan et al. (2010) suggest that a Philanthropic Social Impact Bond will see funds raised from philanthropic sources invested through a special purpose ‘vehicle’ (organization) which would sub-contract with non-governmental organizations to deliver services to achieve the desired outcomes. Thus the SIB arises from the desire to facilitate the ‘private’ (non-government) sector to become involved in the delivery of social services.
Notwithstanding their apparent attraction, reality often complicates the simplest theories and in this article we examine the challenges which might arise in the implementation of PbR mechanisms in the criminal justice system of England and Wales. We start by providing a brief outline of the main attractions of PbR and SIBs to government. We highlight previous and current examples of relevant initiatives in the UK public sector. Next we provide an overview of some of the main challenges to PbR, identifying both conceptual difficulties and more practical challenges. To illustrate these points we use the example of the criminal justice system. We conclude by suggesting that PbR has some limitations and note that there are other innovations in the commissioning of public services which might also address some of the aims of PbR.
Potential advantages of payment by results
A system focused on the delivery of outcomes might look very different to the current system of government provision. It is suggested there will be a wider range of service providers, more variation in models of service delivery and a much reduced role for ‘national standards’. Proponents of PbR and SIBs argue they will drive greater efficiency, innovation and impact in tackling social problems while also reducing risk for government. We examine each of these suggested advantages in turn.
It is suggested, by focusing reward on outcomes, and providing minimal prescription as to how these outcomes should be achieved, PbR models (including the SIB) will drive greater efficiency in tackling social problems. A system that focused on the delivery of outcomes might look very different to the current system with, perhaps, a wider range of service providers, more variation in models of service delivery and a much reduced role for ‘national standards’. 1 One mechanism by which greater efficiency will be achieved is through the correction of misaligned incentives. For instance, an organization responsible for delivering a service has not always shared in the benefits of that activity (Mulgan et al., 2010). For example, a local authority which provides services to divert young people away from crime does not necessarily share in the savings realized by the Criminal Justice Sector arising from reduced prison numbers (Mulgan et al., 2010). A second mechanism by which greater efficiency will be achieved is by allocating resources to where they will achieve the most impact. For example, it is suggested that SIBs are particularly appropriate as a mechanism to focus attention on preventative measures (Mulgan et al., 2010). The recent review of Early Intervention by Allen (2011) provides numerous examples of interventions which fall into this category. Many of these have a long-term objective of reducing offending.
Proponents of PbR argue ‘freeing’ up providers to deliver services in different ways will encourage greater innovation. The focus on outcomes which is a part of the PbR initiative, and the associated reduced focus on commissioners ‘micro-managing’ the processes put in place to achieve those outcomes, seems to be a rejection of the prevailing ‘New Public Management’ approach which has dominated the public sector 2 and, increasingly, the voluntary sector services over recent years.
By making some or all of payment to a service contingent on that service delivering agreed outcomes, PbR transfers risk away from the branch of government commissioning the service and towards the service provider. Government will only pay if outcomes are achieved. The service provider will have to raise some or all of the funds to deliver the service prior to being paid for results and so, from a government perspective, not only is risk transferred to the provider but payments are deferred. Given the need to reduce public sector spending (see above), both the transference of risk and deferring payment for services are attractive propositions for government.
Implicit in the concept of more innovation and the transfer of risk is the potential of PbR models to promote more market testing of public services and encourage new market entrants, particularly from the private and voluntary sectors. The current government has a commitment to ensure that proportions of specific public services should be delivered by independent providers, including civil society organizations, by which it means mutuals, cooperatives, charities and social enterprises. They also have an aspiration to award 25 per cent of government contracts to Small and Medium Sized Enterprises (SMEs) (Cabinet Office, 2011).
Thus, in relation to criminal justice the Ministry of Justice Green Paper suggests that:
The payment by results approach will encourage innovation and bring out the diverse skills from all sectors. We must ensure our commissioning model harnesses the creativity and expertise that independent providers can bring. This includes the small and specialist voluntary providers and social enterprises. (Ministry of Justice, 2010a: 41)
Payment by results in the UK
The UK government is committed to ‘introducing payment by results across public services’ (Cabinet Office, 2011: 9). Currently schemes are being discussed or actively developed in areas including substance misuse, youth justice and children’s services. In this section we discuss some of these.
By far the biggest implementation of a PbR mechanism in the UK must be the Department of Work and Pensions (DWP) Work Programme. The Work Programme is described as ‘the centrepiece of the Government’s plans to reform welfare-to-work provision in the UK’ (Department for Work and Pensions, 2010: 2) 3 The Work Programme brings together various programmes designed to help people back into work into a single scheme. The Work Programme is being delivered by large, private sector providers delivering 40 Work Programme contracts across 18 areas with at least two providers in each area (Department for Work and Pensions, 2010). Most of the ‘prime’ contractors are experienced providers of job-related services ( The Telegraph, 2011) and all have sub-contracted elements of service provision to a range of private and voluntary sector organizations. 4 These contracts will be based on a PbR model with three main payments: an attachment fee paid when a benefit claimant starts on the programme; a job outcome fee paid when claimants enter work; and sustainment payments paid for keeping claimants in work (DWP, 2011).
A range of tariffs has been set for different customer groups, reflecting the need to create incentives for prime contractors to address the increased difficulty of placing some groups in employment (Work and Pensions Committee, 2011). The maximum differential payments for each customer group are set out in the Invitation to Tender document, with job outcome fees varying from £1200 for customers claiming Job Seekers Allowance (JSA) to £3500 for those coming off Incapacity Benefits. Once attachment fees and sustainment fees are taken into account the maximum fee a provider can attract for an individual client ranges from £4050 for a JSA claimant aged 18–24 to £13,120 for an ex-Incapacity Benefit claimant (Work and Pensions Committee, 2011). It is the DWP’s intention that ‘a significant amount of the total amount we’re willing to pay will be in sustainment payments’ (DWP, 2010: 6). It is reported that only 10 per cent of the value of the contract is paid ‘up front’ through attachment fees ( The Telegraph, 2011).
Payment by Results is also being used in other policy areas. One of these is substance misuse. One of the major aims of the 2010 Drugs Strategy, Reducing Demand, Restricting Supply, Building Recovery was ‘to put the goal of recovery for those who are dependent on drugs or alcohol at the heart of all activity’ (HM Government, 2010: 2). As part of this commitment to recovery the Department for Health has launched a pilot of payment by results in eight areas in relation to drug dependency and recovery services. The initiative will aggregate existing funding streams and align overlapping programmes to increase the funds available for providers who are focused on achieving recovery outcomes. Providers will be paid on the achievement of outcomes such as the recipients being free of drug dependence or back in work (Battye and Sunderland, 2011).
Moving to the criminal justice system there are several initiatives which are either direct attempts to implement PbR models or which have established the policy context within which PbR will be implemented. For example, in its Green paper the Ministry of Justice (2010a: 10) states that:
Significant amounts of public money have been spent on rehabilitating criminals without properly holding services to account for the results they achieve. We will move to a new approach where providers are increasingly paid by their results at reducing reoffending.
An example of this approach is the Youth Justice Investment Pathfinder Initiative. This will see a portion of the central youth custody budget invested in ‘pathfinders’ where consortia of local authorities or a single local authority use the funding to commission and deliver their own responses to reduce levels of youth custody and youth re-offending in their area (Youth Justice Board, 2010). Pathfinders will share the financial risks if the custody rate increases and keep the funding if custody numbers are kept low.
Indeed, PbR is seen as a key element of criminal justice reform:
The principle of incentivising performance through payment by results, with success based on the absence of re-offending, should be introduced for prisons, the providers of community sentences and the providers of rehabilitation programmes – whether in the public, private or voluntary sector. With devolved responsibilities and new incentives, we can create a revolution in how offenders are managed, and drive down re-offending. (Conservative Party, 2009: 49)
The Coalition Government proposes that rehabilitation services will be paid for in part according to how successful they are at reducing re-offending; a basic tariff will be paid to cover their costs and an additional tariff paid if targets to reduce re-offending are met. In delivering this they quote the ‘Peterborough Social Impact Bond’ as a promising model (Clarke, 2010).
The Social Impact Bond initiative recently begun at HMP Peterborough is the first pilot of models of this kind in the English and Welsh criminal justice system. The Ministry of Justice has signed a contract with Social Finance (2010) to attempt to reduce the re-offending of a cohort of 3000 adult males who are discharged from HMP Peterborough having served sentences of less than 12 months in custody. The outcome measurement framework is designed to reduce the likelihood of ‘cherry-picking’ whereby the service provider concentrates on clients more amenable to change. Thus outcomes are measured for all offenders discharged from the prison, rather than just those who engage with SIB-funded services and frequency of reconviction is measured rather than a binary measure of whether offenders were reconvicted or not (Disley et al., 2011: iv).
There is evidence that the Peterborough SIB has introduced some new funding into the delivery of criminal justice services, thus transferring risk away from government to non-governmental investors. 5 Disley et al. (2011) report that investors 6 have put £5m in social impact bonds to fund the rehabilitation work and that they could earn a return of up to £8m from the government and the Big Lottery Fund if re-offending among three cohorts, each of a thousand offenders falls by 10 per cent or if the rate of re-offending for all 3000 offenders falls by at least 7.5 per cent. If a reduction in re-offending beyond 7.5 per cent is delivered investors receive an increasing return capped at 13 per cent over an eight year period (Social Finance, 2011:3). The government might realize substantial savings if the scheme pays off. Conversely, if offending does not fall, the government has little down-side risk, while investors potentially lose all their money.
Methodological challenges
Analysis of the underlying concepts behind PbR together with early experiences from approaches being piloted in the UK suggests a number of challenges which must be faced if PbR mechanisms are to work effectively. Here, we outline three methodological challenges, before going on to list the more practical challenges facing the implementation of PbR projects.
The first challenge is that of ‘gaming’. A problem of all performance management systems is that they encourage service providers to concentrate on achieving those outputs and/or outcomes that are included in the system. In doing so, service providers might ignore or neglect other measures even though they know that these are also important. The aim of government in replacing specific targets with more generalized outcome measures is that this problem will be avoided (Hoverstadt, 2011). However, it is not clear that this scenario will be avoided, as Hoverstadt (2011: 1) arguing from a systems thinking perspective explains:
… merely shifting the level of abstraction of what is being measured upwards from activities to ‘outcomes’, doesn’t alter the systemic structure and the same unfortunate consequences are likely to ensue.
In this sense, PBR is no different from any other performance management system and while outcome measures may be appropriate at a governmental level, the challenge comes when they are implemented and providers assume that they are the only measures that count (Hoverstadt, 2011). In this scenario ‘gaming’ is likely and in particular ‘cherry-picking’ whereby providers might select ‘easy’ cases to work with and ignore ‘hard’ cases; thereby maximizing their returns without accepting fully the transfer of risk envisaged by the public sector commissioner (Battye and Sunderland, 2011). For example, in their review of the Work Programme, the Committee for Work and Pensions notes that:
Previous contracted employment programmes have experienced ‘creaming and parking’, whereby providers focus their attention on the participants who are most likely to gain sustainable employment, at the expense of those who face greater challenges to finding work. (Work and Pensions Committee, 2011: 3)
This issue can, in part by ameliorated by giving providers higher payments to work with more challenging cases. So, for example as Battye and Sunderland (2011) describe in relation to the Work Programme (described above) higher payments will be received by providers who are able to place harder to reach cases into sustained jobs. However, as we argue below, identifying the correct tariff to incentivize the provider to work with hard cases is not always straightforward.
Secondly, the challenge of measuring outcomes is a substantial one, particularly in sectors where outcomes are difficult to define and evaluate. Early findings from the Peterborough SIB suggest that:
The development of a methodologically robust outcome measure, which had the confidence of all stakeholders, was a time-consuming and analytically complex process. (Disley et al., 2011: iii)
Fox and Albertson (2011) discuss this challenge at length and consider three main issues: defining outcomes; scale of change; and evaluation methodology. We consider each of these, briefly below.
i. Defining outcomes
Key to PbR models is a clear definition of the outcome which is being purchased. Using the example of criminal justice, in the English and Welsh criminal justice sector this is not straightforward. Consider, for example, the seemingly straightforward result of reducing re-offending: re-offending will generally be measured via the proxy of re-convictions (i.e. ‘proven re-offending’). Thus, it would appear to matter little to the PbR model if offenders cease to offend, or merely become less easy to catch. Further, even this proxy can be defined differently. As noted in the recent Ministry of Justice (2010b) consultation on offender statistics, at present, there are six different measures of re-offending. If outcomes are difficult to define and measure, the result is likely to be that PbR models in the criminal justice system will tend to concentrate on the narrow set of outcomes which is easiest to define and measure.
ii. Realizable scale of change
Standard economic theory suggests that a provider signing up to payment by results will want a return on investment that reflects the increased level of risk assigned to them, and the investor in a SIB will wish to be further compensated for their taking on a risk which is arguably higher than that available in the regular bond market. Mulgan et al. (2010) argue prudent impacts for programmes seeking reductions in re-offending would be in the range of 10 to 20 per cent. However, evidence from criminal justice evaluations indicates even well-executed projects and programmes often bring about only relatively small changes; a point made by Berman and Fox (2010) in their thought-provoking examination of success and failure in criminal justice policy reform. If only small changes to outcomes are possible, this potentially creates problems for investors looking for a decent return.
iii. Evaluation methodology
Key to PbR is the ability of a commissioner to be confident that the desired outcome has been achieved and that it was achieved because of the actions of the service provider that has been commissioned. Thus evaluation designs with high levels of internal validity (experiments and quasi-experiments) will be preferred to those with lower levels of internal validity. If we adopt this approach to the evaluation of PbR models, the result, as Mulgan et al. (2010) suggest, will be the widespread use of evaluations which make use of comparator groups and this will raise the overhead costs of PbR substantially. 7 That is, PbR and SIB investors will be required to raise funds to cover the intervention, and a further tranche of funds to cover a costly evaluation.
PbR, it is said, is likely to foster innovation in social service delivery. Thus extensive evaluation both of outcomes and implementation in different contexts will be required. Every time an innovative intervention is implemented, there will be an ongoing need for further objective evaluation. Overhead costs are therefore likely to remain high. However, detecting small changes in outcomes measures will require large intervention cohorts in order to show statistically significant results (Mulgan et al., 2010). This will limit the range of interventions offered, and the organizations able to become involved.
The third methodological challenge comes when we consider how fiscal benefits to public agencies might be identified and allocated. Government argues that payment by results, as part of a broader approach to change the way public services are commissioned and to promote localism, must be based on a ‘full understanding of the value of the potential results’ (Cabinet Office, 2011: 18). By this they mean:
Understandings of value should be driven by citizens and communities, wherever appropriate. This will enable commissioners to focus services on the social, environmental and economic priorities of the people they serve. (Cabinet Office, 2011: 18)
In economic terms this would imply that the viewpoint to adopt when valuing outcomes in a payment by results contract should be a broad one in which the full social value is considered. But this is also one of the main paradoxes of the payment by results approach. If PbR models are to become widespread they will rely on the commissioning organization being able to realize a monetary saving from which to pay for the results. This implies taking a relatively narrow view of value, one in which value is limited to the fiscal benefits realized by the commissioning organization over the lifetime of the PbR contract.
Consider the example of a PbR contract in which the outcome is a reduction in (criminal) re-offending. Work by economists at the UK Home Office has produced a set of valuations of crime (Home Office, 2005) which could be used to place a value on a reduction in re-offending. However, half of the costs of crime identified are a mix of public (fiscal) and private sector costs. If re-offending declines it is not clear these costs can be ‘saved’. Savings in private sector costs cannot be recovered by a public sector commissioning body. The remaining cost savings are the physical and emotional impact of reduced victimization (Home Office, 2005: Chart 4.2). These are not costs which can be recovered by the commissioning body as we cannot identify those who would have been victims had re-offending not declined.
It should further be borne in mind that not all potential fiscal savings can, in practice, be converted into cash. Not all such costs are marginal – that is, they will not decline proportionally as re-conviction rates fall. The intervention impact which might allow a whole wing of a prison to close might well be beyond the scope of a single PbR model. As Clear and Austin (2009: 315) argue ‘rehabilitation programs, no matter how good, cannot by themselves truly substantially reduce incarceration rates by reducing recidivism’.
In any event, there is no guarantee the desistence from offending of one individual might save any resources. There might, in effect, be a waiting list for prison places in a supply-constrained equilibrium. Roman (2004) shows how it is necessary to understand whether demand for a service is finite.
The Peterborough Social Bond, which is often cited by the government as an example of payment by results in operation avoids these issues because the bond is underwritten by the national Ministry of Justice and not local commissioners. The bond appears to be based only on broad economic returns to the Ministry of Justice which would result from a reduction in re-offending, rather than specific fiscal savings local commissioners would have required. The evaluators of the Peterborough SIB conclude that ‘The Peterborough SIB is too small to deliver substantial ‘cashable’ savings (monetized benefits)’ (Disley et al., 2011: iv). Their suggestion is that to find out whether cashable savings can be identified the SIB model would need to be implemented on a larger scale (Disley et al., 2011).
The evaluators of the Peterborough SIB found that development of a payment model required ‘considerable analytical resources and relied upon the availability of Ministry of Justice data about the cost of reconviction events’ (Disley et al., 2011: iv). By way of comparison, consider the DWP’s ‘Work Programme’. While not without its own complexities, relevant welfare claimants can be numbered in the tens and hundreds of thousands in each region or sub-region. The key outcome – of getting a welfare claimant into work and off welfare is – relatively easy to conceptualize and measure and generates a quantifiable fiscal benefit for government (a saving in welfare payments. 8 Although some of government’s provision for welfare claimants is fixed and would require large numbers of claimants to move into work before a marginal cost could be realized (e.g. the infrastructure to administer benefits or job centres providing support to the unemployed) the welfare payments themselves are a marginal cost and a saving can be made even if the employment status of only one individual changes. (See Mulgan et al., 2010 for further discussion of this contrast.)
In general, it is not clear that monetized savings will accrue in the criminal justice system even where there is an intervention which reduces societal costs. Conversely, if it is the government’s aim simply to reduce the cost of the criminal justice system, the cost of crime to society might well increase. It is, in any event, unlikely that non-monetized savings will be taken into account when defining ‘success’ of a PbR programme.
Implementation challenges
Those involved in setting up PbR schemes are also encountering a number of more practical challenges. Generally, these add to the complexity of establishing PbR and increase the risk of establishing perverse incentives which might encourage intervention providers away from achieving the socially efficient outcomes towards a more tightly defined ‘target’.
a. Individuals versus groups
As Battye and Sunderland (2011) note, a key choice for those setting up payment by results schemes is whether to pay for the results achieved for individuals or on the results for a group. They contrast the Work Programme, which will pay for individuals who find and then sustain employment, with the Peterborough Social Impact Bond which will pay on the basis of re-offending rates for specified cohorts. When payments are based on the performance of individuals one challenge will be to develop a monitoring and payment system that is not overly complex and bureaucratic and runs the risk of substantially increasing the costs of administering payment by results. When payments are based on group performance targeted populations will need to be of a sufficient size to allow statistically robust conclusions to be drawn. So, for example, Disley et al. (2011) describe how, for the Peterborough SIB, the size of impact required to trigger payment was driven, in part, by working out what size of impact could be identified in a cohort such that the finding would be statistically significant.
b. Segmentation
A balance will need to be struck between segmenting populations enough to be confident that services will be tailored sufficiently to the needs of particular groups (Battye and Sunderland, 2011) and defining groups that are large enough for statistically robust conclusions to be drawn about any outcomes achieved (see above). It is likely that the process of assessing client needs and allocating them to different segments will need to be undertaken by a different organization to the one offering treatment so as to avoid the potential for gaming.
c. Setting the tariff
Battye and Sunderland (2011) suggest two broad approaches to setting the tariff paid for achieving outcomes for particular client groups. One approach takes the average current cost of service delivery and has the advantage of being at a level where the commissioners can be confident that providers can deliver but has the possible disadvantage of ‘locking in’ any current inefficiencies (Battye and Sunderland, 2011). The other approach estimates the economic value of the outcomes avoided. Battye and Sunderland (2011) suggest that this is an analytically more sophisticated approach that gives the commissioners a clearer sense of what they ought to be willing to pay for a given outcome. However, the potential drawbacks they identify include the creation of perverse incentives such as disruption to (and even loss of) current providers who may not be equipped to make the transition to PbR.
d. Risk and the balance of service providers
Part of the aim of a payment by results model is to transfer risk from the commissioner of a service to the provider. However, according to the proportion of payments that are dependent upon outcomes and the length of time taken for outcomes to be achieved, one potential drawback in doing so is that smaller providers may be unable to operate because of the pressure this places on their cash flow (Battye and Sunderland, 2011). Smaller organizations are less able to provide a diverse portfolio of interventions than larger providers. In this situation, the failure of one project to meet its PbR target may wipe out the profit of the entire portfolio. A larger organization will be more able to absorb financial losses on one intervention by returns on another. Thus the risk undertaken by the service provider is proportionally greater the smaller is the provider and the more specialized is the provision.
In some cases organizations delivering services under a payment by results contract may need to raise working capital; this will present its own challenges, particularly for smaller organizations not used to doing so and potentially also for some larger public and not-for-profit organizations whose constitutions make such an option either difficult or impossible.
The underlying difficulty here then is that payment by results, rather than leading to a greater diversification of providers and new opportunities for voluntary sector providers and Small and Medium Sized Enterprises (SMEs) actually might lead to the opposite and exclude smaller organizations and voluntary and public sector organizations from the provision of services. Such an outcome would run counter to the aspirations set out in the Cabinet Office (2011) Green Paper Modernising Commissioning: Increasing the Role of Charities, Social Enterprises, Mutuals and Cooperatives in Public Service Delivery.
e. Allocating payment between organizations
Offenders are often a client group with complex needs, particularly offenders serving sentences of less than 12 months (National Audit Office, 2010) where achieving an outcome such as a reduction in re-offending might require several service providers to work together. Similarly, in the field of substance misuse the ‘recovery journey’ will almost certainly involve inputs from a number of providers specializing in different aspects of treatment, recovery and rehabilitation. The challenge this raises is how payments for outcomes will be apportioned between providers, given that the outcomes achieved by their respective services might be contingent on the performance of other service providers.
Commissioners might address this issue in one of two ways. One option is for commissioners to contract with a ‘prime’ contractor and expect them to negotiate sub-contracts with smaller, more specialist service providers able to address particular needs or engage effectively with particular segments of the client group. This is the model adopted by the DWP (described above) for the Work Programme. A second option is for commissioners to enter into separate contracts with a range of providers. This seems to be an approach that might be adopted in some of the drug services payment by results pilots. However, it adds a layer of costly complexity and negotiation.
f. Carrying payments over financial years
Commissioning in the UK public sector normally follows a cycle determined by established financial years. Public bodies that seek to deliver services through a PbR contract may struggle if they cannot carry funds over from one financial year to another and hence ‘smooth’ cash flow. Thus, where PbR contracts envisage longer-term outcomes, commissioners will have to establish procedures that allow funds to be carried over financial years and ‘ring-fenced’ in the expectation they may be required to meet future liabilities. Public bodies wishing to participate, but which are prohibited by law from carrying forward funds, will have to form ‘Special Purpose Vehicles’ in order to deliver services under PbR. However, this latter option will present its own challenges.
g. ‘Special Purpose Vehicles’
Organizations from the public, private or not-for-profit sectors wishing to deliver contracts under PbR may seek to organize themselves through the use of a Special Purpose Vehicle (SPV). A SPV is a legal entity (usually a limited company) created to fulfil specific objectives (Mulgan et al., 2010). A SPV might help a public body circumvent restrictions imposed by central government on carrying over funds from one financial year to another. However, an SPV might also have other advantages:
it might allow a body otherwise restricted from doing so to raise finance in the open market;
it might provide a mechanism to allow a number of organizations to work together to deliver a new service; and
if established as a limited company, the risk associated with a PbR project will be contained by the SPV and not spread to the organization(s) which set it up.
However, establishing a SPV might also present challenges. For example, staff transferred into such an entity will be subject to the Transfer of Undertakings (Protection of Employment) Regulations. For some public sector bodies with substantial pension liabilities ‘TUPEing’ staff into a new SPV might raise difficult questions about future pension obligations. Creation of an SPV might also raise transaction costs and overheads for the delivery of PbR (Mulgan et al., 2010).
h. Settlements with providers
Commissioning cycles in the UK will be of several years duration. Thus, a provider will bid to deliver a service for an agreed number of years and, at the end of that period, the service will be re-commissioned. Normally, if the previous provider does not win the contract to continue to deliver the service a ‘clean break’ between the commissioner and provider is possible. The same is true if either side use a ‘break clause’ to end a contract for the provision of services. However, under payment by results achieving a clean break will be much more difficult. At any given point in the commissioning cycle there will be future liabilities for the outcomes achieved by a cohort currently being served. Monitoring, and potentially evaluation, may need to continue after the life of the current commissioning round before final payments based on longer-term outcomes can be agreed. The alternative for parties wishing to make a clean-break will be to agree the potential future value of the contract and a one-off payment to cover all future outcomes delivered and liabilities incurred. Working out the value of such a payment will be difficult and the potential for legal action high.
Discussion
Payment by results models do have a place in the provision of ‘public’ services. The transference of risk is attractive to government. In some parts of the public sector, where the potential client group is fairly large and homogenous and service provision is highly integrated, measurement of outcomes is straightforward and the fiscal return on achievement of outcomes is easy to establish. Here, a payment by results mechanism might be attractive to commissioners, service providers and investors looking for a return.
However, as we illustrated through a consideration of the potential for payment by results in the criminal justice sector, where outcomes are relatively difficult to measure and value, service provision is complex, fiscal benefits are shared across a number of potential commissioners and existing evidence on what works is still relatively patchy, payment by results is unlikely to be a panacea – either as a driver of service reform, or as a means of attracting new sources of funding into the sector.
There is an inherent contradiction in PbR. In general, private investors have a higher discounting factor than government, that is, they expect to see their investments yield returns in the short-term, yet the social return on any social investment is likely to be long-term. There are few quick-fixes in criminal justice. In addition, investors will be looking for a return which covers not only the cost of the intervention, but also the cost of the evaluation. Finally, thinking particularly about SIB, given that the government is limiting its downside risk, while investors will lose their capital if their projects do not pay-off, only those projects likely to lead to substantial benefits will be offered – those projects where social returns are either large, certain or both. This begs the question, why doesn’t the government invest in such projects already on behalf of the taxpayer?
Given these uncertainties it is therefore worth returning to the aims of the PBR experiment and asking whether there are other innovative approaches to commissioning public services that might deliver some or all of the same aims. As set out earlier in this article, proponents of payment by results mechanisms suggest some combination of four main advantages: greater efficiency; more innovation; the transfer of risk away from commissioners; and the encouragement of new market entrants. Our ongoing work on criminal justice reform suggests that two alternative approaches might be particularly relevant in England and Wales.
The first is a relatively well-developed concept, particularly in the United States; that of Justice Reinvestment (Tucker and Cadora, 2003). As Fox et al. (2011) explain, Justice Reinvestment seeks to reduce the level of crime in the most efficient way possible, potentially creating a more law-abiding society at a lower cost than the traditional detect/convict/punish approach. Clearly the crimes which cost society the least are those which are discouraged before they are committed. Hence, Justice Reinvestment (JR) proposes moving funds spent on punishment of offenders to programmes designed to tackle the underlying problems which gave rise to the criminal behaviour (Allen, 2007).
Justice Reinvestment has the potential to deliver several of the benefits of payment by results. This is clearly an innovative approach that promises not just the development of new intervention services, but a change in the culture of the criminal justice system. Through a greater focus on tackling criminality ‘up-stream’, resources should be used more efficiently. However, this will be a long-term process and so efficiencies will take time to be realized. The greater emphasis in Justice Reinvestment on recognizing that offending behaviour has its roots in families and communities also holds out the possibility of greater diversity of service providers. However, Justice Reinvestment is politically less appealing at a time of austerity measures. Justice Reinvestment requires up-front investment in the first instance. Government budgetary savings are being looked for now, while savings resulting from efficiency and social innovation will all take time; perhaps a time-scale beyond the electoral cycle.
It has also been argued, by Clear (2011), that Justice Reinvestment suffers from some of the drawbacks which we contend are inherent in the PbR and SIB approach. If funds are to be reinvested, we must have a rigorous evaluation of where savings have been made and of the magnitude. Clear also contends the original vision of Justice Reinvestment, i.e. strengthening communities, has not always been realized. He argues (Clear, 2011: 593) that in the USA ‘evidence-based strategies often have resulted in redirecting government funds from corrections to other government-based social services’. This is in contrast to the original vision of Tucker and Cadora (2003: 1) who suggest Justice Reinvestment ‘is also about devolving accountability and responsibility to the local level. Justice reinvestment seeks community level solutions to community level problems’.
A second possibility that draws on experience in the social care sector, but that has yet to be implemented fully in the criminal justice sector is personalization. Personalization has been rolled out widely in the social care sector. In a personal budget based system, people who have been assessed (or in some cases, have self-assessed) as eligible for a service are told how much money is available to fund their service and are given the option of taking control of that money, either through taking a cash Direct Payment, or through co-producing a spending plan with the council (a managed personal budget), an independent brokerage organization, or a service provider (an Individual Service Fund) (Department of Health, 2010). Personal budgets and Direct Payments have been transformative for a significant proportion of people using care and support services although it is worth noting that they were only envisaged to be one element of a wider reform. The ministerial concordat Putting People First (Department of Health, 2007; but see also Crosby and Duffy, 2008) set out a comprehensive vision for personalizing social care, across four quadrants:
a universal offer of advice and information to help people make informed choices;
a focus on developing inclusive and supportive communities (‘social capital’);
a focus on investing in prevention;
introducing choice and control through the introduction of personal budgets and self-directed support.
To date progress on the other three quadrants has been slower, but in relation to personal budgets most service users and their carers report that taking a personal budget has had a positive impact upon their lives (In Control, 2011). There are some examples of a more plural and creative market for services, including a large number of voluntary sector providers and a growing number of non-traditional approaches and enterprises (Fox et al., forthcoming).
Over recent years there has been an emerging dialogue in offender rehabilitation about more individualized services. This has been evident in research and theorizing around the concept of desistance (McNeill, 2009) and the emergence of the good lives model of offender rehabilitation (Ward and Maruna, 2007). Concepts such as ‘end-to-end’ management have placed more emphasis on continuity in the relationship between an ‘offender manager’ and the offender, and on developing a package of services that are tailored to each offender’s needs. Fox et al. (forthcoming) argue that drawing on the experience of personalization in the social care sector would bring three important new elements to work to reduce the likelihood of an offender re-offending. First, co-production offers new opportunities for offenders to be empowered to take ‘ownership’ of factors that have contributed to their offending behaviour and to develop their self-reliance and personal capacity. The social care experience suggests that previously excluded people can become full citizens when they have access not only to rights, but also to responsibilities. Similar advantages might be felt by offenders. The developing literature on desistance stresses the importance of ‘agency’ in the offender desistance process and already some commentators are pointing to the concept of ‘co-production’ as a potential approach to adopt in the criminal justice system (Weaver, 2011). Secondly, the use of personalized budgets could stimulate innovation in the criminal justice sector by putting in place opportunities for the development of new providers and new, more specialist provision in the criminal justice sector. Thirdly, individual support which had a focus upon people’s ability to form and sustain informal relationships, alongside community development work, could help offenders establish more positive and socially-included patterns of living.
The concept of Justice Reinvestment and personalization might be brought together using the portfolio of investment approach of the Washington State Institute for Public Policy (WSIPP). Their sophisticated economic modelling project combined analysis of the offending population in Washington State, a wide-ranging systematic review of the evidence-base on what is effective in reducing crime and re-offending, together with long-term economic modelling (Aos et al., 2006). This project has allowed WSIPP to develop a number of long-run economic models for different justice reinvestment strategies characterized according to how ‘aggressive’ investment is in diversionary measures.
Such information on ‘what works’ might inform offenders and offender managers in co-producing personalized resettlement plans. Community Budgets could provide a financial mechanism for pooling relevant funding streams. 9 Local Community Safety Partnerships provide strategic direction to the development and coordination of appropriate, geographically-targeted services.
Conclusion
Payment by Results is, on the face of it, an attractive concept. Commissioning services to deliver key social outcomes and freeing up professionals to decide how best to deliver those outcomes seems sensible, and, in the current financial climate, who would argue against approaches that encourage greater efficiency and more innovation? However, it is not clear to us that Payment by Results is the only or most effective way to deliver these outcomes. The Social Impact Bonds which, it is suggested, may fund Payment by Results interventions, might bring new funds into the delivery of public services, but it is in this area that PbR seems most uncertain.
In sum, the stated aims of social interventions delivered through Payment by Results appear to be optimistic. Although this approach appears to be firmly founded on simple theoretical models, there are inherent flaws in application which will, we believe, limit its use and the benefits which might be gained from it. It is not altogether clear that society as a whole will benefit from their introduction and, if philanthropic entrepreneurs are not forthcoming, we are concerned services might suffer. With careful commissioning and monitoring, SIB funded PbR interventions might provide a part of the answer to reducing costs while increasing performance. They are unlikely, however, to provide the full solution.
Therefore, while it is sensible for the government to pilot PbR in the criminal justice system, close analysis of the concept and early evidence from its implementation suggests that the government would be wise to be cautious in assessing the potential for PbR. While we wait for the results of PbR pilots, the government should continue to explore other innovative approaches to transforming the criminal justice system.
We suggest, where interventions are proven to be efficacious, there is no reason why government should not invest in such on behalf of the taxpayer, rather than tender for the private sector to do so. This is not to say that the private sector and voluntary sector might not also offer social innovations of benefit to investors and society; innovations which may be more efficiently delivered at a local level. In such circumstances PbR and SIBs may still have a (modest) role to play. They may motivate and support the provision of a complementary intervention to government social services, rather than a substitute.
As Tucker and Cadora (2003: 4) put it in their seminal paper setting out an early vision for Justice Reinvestment: an offender’s
[r]eentry must be a geographically targeted partnership of public and private interests—penal, social services, health providers, and educational institutions. No [one] size fits all.
By utilizing PbR, personalized budgets and the portfolio of investment approach of the WSIPP, we might, just might (with good will, evidence based practice and a genuine desire for improvement in the lot of poorer communities) be able to get the best of both worlds; private and public.