Supporting moves into work: New Deal for Disabled People ﬁndings

New Deal for Disabled People (NDDP) is a voluntary programme designed to help people with disabilities and health conditions secure (sustainable) employment. It was delivered through a national network of public, private and voluntary sector organizations (known as Job Brokers). This paper discusses the impact of NDDP and how its relative success was enabled by the wider institutional environment. NDDP was cost-beneficial in both reducing incapa-city-related benefit receipt and increasing the employment rate of programme participants. The institutional factors covered are: programme take-up, contract management and funding regime, and Job Brokers relationships with the public employment service, Jobcentre A version of this was originally prepared for a Panel on Disability Employment Policy: Findings from Three Path-Breaking UK Studies and Their Lessons for the US at the Association of Policy Analysis and Management Conference What Else Shapes Public Policy Analysis and Management? 8


Introduction
In recent years policymakers in the UK and elsewhere have implemented programmes to help people with health conditions and disabilities move into employment in order to reduce benefit rolls. The UK Government piloted New Deal for Disabled People (NDDP) in 1998, and implemented a revised version nationally in 2001. It is a voluntary programme whose delivery has been contracted-out by the Government's employment and benefits service, Jobcentre Plus, to a national network of providers known as Job Brokers. The programme is designed to help people claiming incapacityrelated benefits move into sustainable employment. An evaluation of NDDP commissioned by the Government's Department for Work and Pension shows that the programme has been a 'success' in the narrow sense that more participants than non-participants left incapacity-related benefits and entered work, and from a societal point of view the benefits of the programme exceeded its costs. 1 The findings reported here are based upon this evaluation which was conducted by a UK and US consortium independently of government.
NDDP is of interest, firstly, because as an employment programme it is unusual in that it provides a framework consisting of a funding and contractual regime and a sought after number of (sustainable) job outcomes rather than a programmatic theory of how to secure job entries for disabled people. The Government did not prescribe in advance the types of service that Job Brokers should provide in order to assist disabled people into employment. There is no theory contained within the Government's policy that causally links objectives with outcomes. The 'vessel-like' quality of the programme means that responsibility for conceiving any programmatic theory for informing practice was effectively devolved to providers. Secondly, NDDP is one of the few UK welfare to work programmes where all of the provision has been contracted-out. It provides an example of the mixed economy of welfare as it applies to disabled people and employment programmes.
This paper highlights NDDPs impacts, costs and benefits, and the main institutional factors underpinning its success. It argues that two factors, the broader institutional framework within which the contracted-out programme operated and the funding and contractual regime adopted, have had a key role in the 'success' of the programme. Before discussing these issues some further details about NDDP and the participants on the programme are presented.

Policy background
The Labour Government, elected in 1997, was from the outset intent on reforming welfare provision and reducing benefit dependency (Cm. 1998), and in particular determined to encourage disabled people to return to employment (DSS 1998). Ministers were concerned about the numbers claiming incapacity-related benefits (2.5m or 7% of the working age population in 1998 (DWP 2005)), the associated public expenditure implications (£7.2 billion in 1998/1999on Incapacity Benefit (1998/1999 (DSS 2000)), and that around one million of the client group claimed to want to return to work (see Stafford 2003Stafford , 2005. NDDP was piloted in September 1998 and subsequently extended nationally in July 2001. (There are differences in the design of NDDP as piloted and as implemented nationally (see Stafford et al. 2007) and this paper focuses on the nationally extended version of the programme. Further details about the pilots can be found in Arthur et al. (1999), Blackburn, Child, and Hills (1999), Hills et al. (2001) and Loumidis et al. (2001).) NDDP was available to people in receipt of certain qualifying benefits, notably in 2001 the 1.5m recipients of a social insurance benefit, Incapacity Benefit (DWP 2005). Between 2001 and 2008 NDDP was delivered by approximately 60 Job Brokers, who were a mix of public, private and not for profit organizations who competitively bid to deliver the programme. The precise mix of services offered by individual Job Brokers varied, but in general Job Brokers offered tried and tested methods for helping the client group secure employment: assessment of basic English and mathematical skills; production of individualized Action Plans; advice about vocational direction; advice and help with job search; financial advice; in-work support; and training and work placements.
In addition to contacting Job Brokers directly, new incapacity-related benefit claimants could access NDDP through approximately 1000 local Jobcentre Plus offices (NAO 2008, 8). A key feature of the Jobcentre Plus model is the mandatory work-focused interview, which takes place at a local Jobcentre Plus office, and involves personal advisers explaining Jobcentre Plus services, identifying customers' barriers to work and outlining the help and support available to them.
NDDP does not operate in isolation. Related policy initiatives include the establishment of the Disability Rights Commission (which has now become part of the Equality and Human Rights Commission) and the strengthening of disability anti-discrimination legislation (especially in the workplace), the introduction of the National Minimum Wage, a tax credit for people in low-paying jobs, and other benefit changes to tackle work disincentives (DSS 1998). In addition, incapacity-related benefit recipients could access a range of Jobcentre Plus disability employment services provided by Disability Employment Advisers. One initiative with particular salience for NDDP is Pathways to Work. This national programme offers a package of support for incapacity-related benefit recipients that includes mandatory work-focused interviews; interventions, known as the Choices package, to support a return to work which can include NDDP; and a work incentive, the Return to Work Credit. Indeed, since March 2008 NDDP has been delivered in only certain parts of the UK where provision of Pathways to Work has not yet been tendered to private contractors. Elsewhere similar services are made available by providers contracted to deliver Pathways to Work (DWP 2007). However, the results presented below cover a period when Pathways to Work was being piloted; so some of the findings relate to NDDP-only areas and some to locations where Pathways to Work was being piloted.

Who participated in NDDP?
Participants in the programme differed in certain respects from the wider eligible population; they were more likely to be male, more likely to be younger, and on benefits for a shorter period of time than the Incapacity Benefit population as a whole (Stafford et al. 2007, 46Á50). In addition, participants were less likely to have a mental health condition, but more likely to have a musculo-skeletal condition than the wider population claiming Incapacity Benefit; albeit both conditions predominate in both groups. Moreover, participants, in contrast to members of the eligible population, were in certain respects 'closer' to the labour market, for example, participants were more likely to say their health was fair, good or very good, and less likely that it was bad or very bad; to possess an educational qualification; and to be in, or looking for, work. Notwithstanding these compositional differences, which are taken into account in the impact analysis, NDDP has a positive effect on reducing benefit receipt.
Outcomes, impacts and cost-benefit analysis Outcomes NDDP was effective in both reducing incapacity-related benefit receipt and increasing the employment rate of programme participants. Administrative data show that of the 260,330 registrations on NDDP between July 2001 andNovember 2006, 110,950 (or 43%) had resulted in jobs by November 2006 (Stafford et al. 2007, 111). The proportion of participants finding work also increased over time. So, for example, of those registering between July 2001 and June 2002, 32% found work within 12 months, compared to 44% of those registering between December 2004 and November 2005.
There is also a variation in the proportion of participants entering employment by type of area. The proportion entering work in the Pathways to Work areas (39%) was considerably higher than elsewhere (29%) (Legge et al. 2006, 148 and 150). It is conceivable that this area difference arises from the nature of the interventions available in Pathways and non-Pathways areas, in particular the availability of a Return to Work Credit (worth £40 per week for up to 52 weeks for participants and others with earnings less than £15,000) in the (then) pilot areas. However, the higher employment rate for Pathways to Work pilot areas might also reflect other factors, such as differences in the performance of Job Brokers.
Nine out of ten (93%) participants obtaining a job are employees, with most in full-time employment, and the remainder in self-employment, again most working full-time (Stafford et al., 2007, 128). Participants enter employment in all the major occupational groups, although significantly more, around a quarter, enter routine, unskilled occupations than any other occupation (Kazimirski et al. 2005, 87Á9;Legge et al. 2006, 165Á8). A further one in six participants enter sales and customer service jobs. Jobs also tend to pay above the National Minimum Wage; the mean pay per hour for those working more than 16 hours per week was £6.30 and the mean weekly pay was £215 (in early 2005) (Legge et al. 2006, 160Á5).
A key objective of NDDP is to secure sustainable employment for participants. Administrative data show that of those participants entering work by August 2006, 57% (or 59,080 participants) achieved sustainable employment (defined as employment lasting for 13 or more weeks) (Stafford et al. 2007, 133). This is probably an underestimate as Job Brokers are unlikely to maintain contact with all participants achieving sustainable employment.

Impacts
The impact analysis was conducted using a matched comparison methodology where participants are compared with non-participants. Using administrative data three cohorts of recipients were examined, namely the: Early Cohort Á people who registered between July 2001 and December 2002 and provides estimates of NDDP over a 24-month follow-up period.
Maximum Follow-up Cohort Á for those who registered during the first six months of NDDP this cohort allows impacts to be estimated over a 36-month period; this is a subset of the Early Cohort.
Late Cohort Á covers a later registration period, January 2004 to June 2004, and provides an initial assessment of the programme's effectiveness following the changes to funding introduced in 2003 (see below).
Impacts are assessed for benefit receipt, amount of benefit and employment and separate results are reported for more recent recipients and longer-term claimants.
The impact analysis shows that NDDP has a positive impact on benefit receipt and employment (see Table 1). Analysis of the Early Cohort shows that NDDP is effective in encouraging participants to move off incapacity-related benefits (Orr, Bell, and Lam 2007, 30Á2). By month 24, the reduction in the rate for benefit receipt was 16 percentage points for longer-term participants and 13 percentage points for more recent participants. For the Maximum Follow-up Cohort Table 1 shows that for longer-term participants NDDPs impact on benefit receipt continued to grow, an 18 percentage point reduction by month 36 (Orr, Bell, and Lam 2007, 38Á41). However, for more recent recipients estimated net effects diminished over the period from month 24 (13 percentage points) to month 36 (11 percentage points). This is because there was a small net movement of former more recent participants back onto benefit; suggesting that the progress made in the first two years was beginning to decline over time. A comparison of the Early and Late Cohorts shows that the estimated impacts for the Late Cohort are nearly double those of the Early Cohort. In the Late Cohort, NDDP reduced the net benefit receipt rate of those registering in the first six months of 2004 by 15 percentage points for longer-term recipients and by 19 percentage points for more recent recipients at the end of the six month period (Orr, Bell, and Lam 2007, 42Á4). However, the reasons for this increase in impacts are unclear. It might be due to policy changes to the funding regime (see below) or to other factors, such as the programme becoming more effective over time.
Essentially, the impacts of NDDP on the average amount of incapacity-related benefit received mirrored those for benefit receipt, hence, NDDP generates benefit savings. So, for example, the Early Cohort reveals that average monthly benefit savings initially grew and by month 24 were £81 for longer-term claimants and £51 for more recent recipients (see Table 1) (Orr, Bell, and Lam 2007, 30 Source: Based on figures given in Tables A1ÁA7 and A12Á14 in Orr, Bell, and Lam (2007).
The impacts of NDDP on employment are generally consistent with the impacts on benefit receipt although they are smaller (see Table 1) (Orr, Bell, and Lam 2007, 52). Twenty four months after registration, NDDP had increased the employment rate for longer-term participants by 11 percentage points and by seven percentage points for more recent participants (Orr, Bell, and Lam 2007, 53). The Maximum Follow-up Cohort suggests that for longer-term recipients, impacts on the employment rate level off at around 10Á11 percentage points in the third year after registration (Orr, Bell, and Lam 2007, 56Á7). Impacts for more recent recipients were more variable (due to the smaller sample sizes in the cohort), and at the end of the third year were eight percentage points Á similar to that for the Early Cohort in month 24. The six-month follow-up period available for the Late Cohort, reveals that impacts on the employment rate for more recent recipients were very similar to those in the Early Cohort (see Table 1) (Orr, Bell, and Lam 2007, 57Á8). However, for longer-term participants' employment impacts are two to three percentage points larger in the Late Cohort compared to the Early Cohort. Again, whether this improvement in performance reflects the policy changes to NDDP or other factors is unknown.

Cost-benefit analysis
That the impact analysis reveals that the effects of NDDP are favourable does not, of course, mean that society should devote resources to the programme. The costbenefit analysis examines the net benefits (that is, benefits minus costs in 2005 prices)of NDDP. Separate benefit estimates are given for longer-term and more recent participants on a per participant basis because the impact of NDDP on these two sub-groups varies (see above) (Greenberg and Davis 2007). 2 Uncertainties about the reliability of the costs incurred by Job Brokers mean that three approaches are used to estimate programme costs: Costs are based on the actual outcome-related payments paid by the Department for Work and Pensions to Job Brokers. These net benefits are labelled 'actual government' costs or expenditures in Table 2.
Costs are based on those reported by the 19 Job Brokers who took part in the evaluation's cost study. These are labelled the 'upper-bound' costs or estimates in Table 2.
Costs are based on those reported by 17 Job Brokers and the adjusted costs of two Job Brokers. This adjustment was made because of suspected overreporting of these two Job Brokers' actual costs, and is labelled 'lower-bound' costs or estimates in Table 2.
The estimates discussed below are based on Greenberg's and Davis' third approach, which is their preferred cost measure (2007,29,52). The cost-benefit analysis is presented from three perspectives: the Government, the participant, and society as a whole. The societal perspective is the sum of the costs and benefits for NDDP participants and the Government.
From the Government's perspective (and regardless of the cost estimate used) NDDP is cost-beneficial. It reduces the Government's budgetary requirements by £2815 for a typical longer-term participant and by £1016 for a typical more recent participant (see Table 2) (Greenberg and Davis 2007, 82). In terms of the costs of NDDP this is a considerable saving; for each pound the Government spends on NDDP it saves £4.50 for longer-term recipients and £2.26 for more recent claimants in benefit and administrative expenditure.
For participants the net benefits are positive for a typical longer-term participant (at £348), and negative for a typical more recent claimant (at (£155), but fairly modest in both cases (Greenberg and Davis 2007, 91Á2) (see Table 2). The reduction in the income of a typical more recent claimant arises because their increase in earnings is more than offset by reductions in benefits. The positive net benefit received by a typical longer-term claimant is almost entirely attributable to tax credits (wage supplementation) they receive as a result of employment. There is however, some uncertainty surrounding these estimates for participants (Greenberg and Davis 2007, 92Á3), but it seems likely that a typical NDDP participant benefits from taking part in the programme but only to a very modest degree.
Nonetheless, the modest nature of the net benefit raises questions about why participants enter paid work. The evaluation shows that participants are highly motivated towards work. Around three-quarters state 'money' is the main reason for wanting to work (Legge et al. 2006, 131), but employment is more than a financial necessity, often participants value gaining independence, social contacts and simply being occupied and 'out of the house' (Corden et al. 2003, 91). So for some a modest financial return is probably worthwhile.  Table 6.1, Table 7.1 and Table 8.1 in Greenberg and Davis (2007).
NDDP appears to have positive net benefits from a societal perspective. The social net benefits of NDDP are considerably larger for a typical longer-term recipient (£3163) than they are for an average more recent claimant (£861) (see Table 2) (Greenberg and Davis 2007, 103). In other words, there was a return to society of £4.93 for each pound the Government spends on NDDP in serving longer-term recipients and £2.07 per pound for more recent claimants. The source of the difference between the two claimant groups is NDDPs larger impact on employment and, hence, earnings of longer-term participants.

Institutional framework
The success of the programme, however, needs to be contextualized and in particular the institutional factors Á programme take-up, the management of the contract and funding regime and its implications for the programme, and communication and information channels between Job Brokers and Jobcentre Plus that affected performance Á underpinning the success of NDDP are considered in turn below.

Programme take-up
The crucial role that the wider institutional environment had in the success of NDDP, is illustrated by both the overall take-up rate and variations in take-up rates by type of area. There was a steady build-up in the number of registrations to 260,330 by the end of November 2006. As individuals could register, de-register and then re-register on NDDP, the number of people who had registered over the same period was lower at 231,610. Administrative data suggest that the rate of take-up of NDDP amongst those flowing onto the qualifying benefits increased over time, and the overall take-up rate for the year ending May 2006 was 3.1% of the eligible recent claim population. 3 This appears to be a relatively low proportion but Government never intended the programme to be comprehensive and it compares favourably to, say, a similar programme in the United States (Thornton et al. 2006, 9). The take-up rate partly reflects the relatively high level of awareness of the programme: over half (53 to 58%) of the eligible population were aware of NDDP and/or local Job Brokers (Pires et al. 2006, 55). Awareness of NDDP may have been high because of the active role that Jobcentre Plus had in referring new incapacity-related benefit claimants to NDDP and Job Brokers.
The take-up rate of NDDP was also notably higher in the Pathways to Work pilot areas; for the year ending May 2006 it was 5.2% in Pathways areas compared to 2.7 to 2.9% in Jobcentre Plus offices. The higher rate for Pathways to Work areas is likely to be due to the more intensive work focused interview regime for new (and some existing) recipients.

Funding regime
The funding regime for Job Brokers was largely outcome related and evolved over time following criticisms from providers (Corden et al. 2003, 48). Job Brokers receive a small registration fee for each person joining the programme and larger payments for participants entering paid work and for those retaining their jobs for a sustained period. The registration fee for each participant was increased from £100 to £300 in October 2003. Job Brokers receive roughly equal outcome payments for job entries and sustained employment, but the amount paid varies between Job Brokers and was negotiated as part of the contract procurement process with the Department. For example, payments for entry to a full-time job ranged from £500 to £2368 with a mean of £1155 (2001 prices) (Stafford et al. 2004, 67). To help ensure that the programme budget is not exceeded Job Brokers' contracts contain targets for numbers of registrations, job entries and sustained jobs.
When the programme began many Job Brokers focused on registering people and subsequently some providers accumulated a relatively large number of participants who had not entered work. There was concern within the Department for Work and Pensions that some of these individuals were no longer receiving services and that Job Brokers were not helping them secure jobs. Accordingly, the Department introduced a requirement that each Job Broker convert a minimum 25% of all of their registrations to job entries by March 2004. The Department also introduced in April 2005 district and monthly performance profiles for each Job Broker. These profiles were set by Job Brokers themselves, based on the number of outcomes they anticipated achieving over the life of the contract. The monthly profiles were designed to ensure that Job Brokers did not exceed the maximum amount of funding allowed under their contracts (Davis, Pound, and Stafford 2006, 24). Job Brokers who 'over-perform' against their profile can have their activities 'capped' (Davis, Pound, and Stafford 2006, 27Á8). If Job Brokers 'under-perform' (compared to their profile) and subsequently fail to improve their performance, they may be subject to re-profiling. This involves reducing the number of registrations and job entries a Job Broker is expected to attain for each remaining month of the contract. Some providers have had their job entry profiles cut by as much as 50% (Davis, Pound, and Stafford 2006, 31Á2). This subsequently enabled them to achieve a high proportion of their revised profile and their performance subsequently appears to be much improved.
These changes to the contract and funding arrangements mean that Job Brokers tend to become more focused on (potential) participants who were closer to work and who are likely to move into employment relatively quickly (Lewis et al. 2005, 41). Job Brokers had to be confident that they can convert registrations into job entries within a reasonably short period of time; otherwise their reported performance against contracted target can be undermined. Criteria for judging the suitability of participants for help can differ subtly between Job Brokers; however, key factors are the job readiness of the participant and the amount of time and resources required to achieve a job entry (Davis, Pound, and Stafford 2006, 17). There are two policy issues that arise here. First, if external providers are to deliver services to the 'harder to help' then this needs to be reflected in any funding regime. A more sophisticated outcome funding regime is required, such as including payments for intermediate outcomes (Lewis et al. 2005, 169). Secondly, such selective practices can make programmes inequitable in that they can limit access to those who want to work but are not seen as job ready by providers. Policy needs to recognize that with such a diverse client group interventions aimed at both those closer and more distance from the labour market are needed.
The funding regime changes also affected the type of services provided. Individuals requiring more support, because they are further from the labour market, can be referred to other organizations and services, and this includes referring the 'hard to help' back to Jobcentre Plus (Corden et al. 2003, 43). In addition, offering financial inducements or rewards to participants who have entered work or been employed for 13 weeks became more commonplace amongst Job Brokers.
Notwithstanding these financial pressures, not all Job Brokers engage in 'cream skimming', some claim that they have 'open' registration practices and make the service widely available (Lewis et al. 2005, 41).

Relationships between Job Brokers and Jobcentre Plus
Relationships between Job Brokers and local Jobcentre Plus offices are a key element of NDDP. Qualitative research shows that the more successful Job Brokers (in terms of performance) have good working relationships with at least some local Jobcentre Plus offices (Davis, Pound, and Stafford 2006, 33;Lewis et al. 2005, 151). For Job Brokers links with local Jobcentre Plus offices, especially with Disability Employment Advisers, are important because they are a source of potential participants, Job Brokers could use them to access Jobcentre Plus services, and Job Brokers can use local Jobcentre Plus offices as venues to meet with (potential) participants (Corden et al. 2003, 30Á1). However, the development of a working relationship between the two organizations is in some instances problematic. Relationships can be undermined by feelings of suspicion (Corden et al. 2003, 31Á2). Here the development of the relationship is more ad hoc and only one party can be seeking to establish a link. Indeed, initially some Disability Employment Advisers saw Job Brokers as a threat to their work with customers. In general, however, relationships between the two have improved over time due to more inter-organizational communication (Corden et al. 2003, 31;Lewis et al. 2005, 113). The perceived benefits of co-operation are an increased likelihood of each organization achieving its respective targets and provision of a more comprehensive service to the client group. However, Job Brokers vary in the extent to which they 'market' themselves to local Jobcentre Plus staff. Some give it a high priority and promote contacts, others have relatively few contacts (Lewis et al. 2005, 109).

Conclusion and discussion
The impact analysis shows that the programme 'worked' in that it reduced the number of people in receipt of incapacity-related benefits and increased the number entering employment, especially longer-term recipients compared with what would have happened had the programme not been introduced. Moreover, as a relatively low-cost intervention, NDDP offers value for money (at least from the perspectives of the Government and society as a whole). However, this does not mean that NDDP is evidence for adopting a contractual model to the delivery of employment services to disabled people. This paper shows that there are institutional factors that are influential, notably the support that Job Brokers receive from the public employment service, Jobcentre Plus. Programme take-up benefited from Jobcentre Plus actions. Jobcentre Plus advisers had a significant role in referring customers to appropriate Job Brokers. In addition, reflecting the UKs emphasis on disabled people entering employment, the introduction of mandatory work focused interviews for the client group at Jobcentre Plus allowed the agency to signpost potential users to NDDP (Lewis et al. 2005, 171). Indeed, Pires et al. (2006, 101) argue that engagement with NDDP following compulsory work-focused interviews could initiate a process that might not otherwise have happened.
In addition, the governance arrangements for NDDP place a premium on Job Brokers achieving set performance targets. As a consequence the success of NDDP is not so much attributable to the marketization of the programme as to the incentive structure created by a particular funding regime. The use of a 25% minimum requirement and of monthly and district profiling is arguably not only a form of regulation but also an element of top-down control over delivery. Under these governance arrangements, discretion has been removed from Job Brokers and relocated at the centre. In effect, 'performance' becomes 'conformance' in order to ensure financial accountability and the meeting of targets. Put more starkly such a regime seeks to gives 'control and surveillance' to the central authority (Ryan 1999). There is no easy 'trade-off' here; however, policymakers and providers need to recognize that regulatory frameworks are a factor that can influence providers' expectations and behaviours and hence outcomes.