The curse of accountability: Assessing relationships in the delivery of employment services

Abstract

As governments move from being both a funder and provider of human services to a purchaser of services in private sector markets or quasi-markets, ensuring that providers do what they are supposed to do becomes more difficult. Agency theory and stewardship theory have been suggested as ways of overcoming this problem. This article argues that both are inadequate, particularly because they conceptualise the relationship as bilateral (government funding department and service provider), ignoring the role of clients in achieving organisational objectives. Co-production that recognises the role played by clients in the production of employment outcomes can provide a more useful way of thinking about relationships among key actors involved in the provision of employment services.

Introduction

As governments move from being both funders and providers of human services to purchasers of services in private sector markets or quasi-markets, the issue of control – how to ensure providers do what they are supposed to – becomes increasingly important. As in-house expertise declines, governments are faced with the problem of information asymmetry, that is, providers (or sellers of services) know more about what is required to produce the service than government. In the human services sector where the government is purchasing what Brown et al. (2010) define as ‘complex products’, the problem is particularly acute because ‘the cost, quality and quantity parameters of complex products cannot be easily defined or verified’ (p. i42).

Agency theory assumes that individuals and organisations are extrinsically motivated by economic self-interest, and sees the solution to information asymmetry as one of designing controls that will enforce compliance (Knapp et al., 2011: 296–297). Often the result is a situation where the principal (the government funding department) attempts to control the agent (service providers) through detailed, prescriptive contracts combined with frequent monitoring of the agent’s activities. Critics of agency theory argue that its application to human services, such as health, education or social welfare services, will not ensure compliance on the part of agents because in these ‘caring professions’, agents are professionals motivated not by economic self-interest but by professional norms and values (Broadbent et al., 1996). This also applies to many organisations providing employment services (Nevile and Nevile, 2003).

Stewardship theory is put forward as an alternative framework for conceptualising the relationship between governments that fund services and the organisations that deliver them. Stewardship theory assumes that service providers are intrinsically motivated stewards who value co-operative behaviour over self-interested behaviour and who act to maximise organisational objectives. Consequently, principals need to empower stewards, rather than seek to control them, to ensure high-quality services (Davis et al., 1997: 24–25). However, stewardship theory has also been criticised for failing to take into account the risk-averse environment in which government departments operate (Van Slyke, 2007: 183), and for largely ignoring both the social context within which agents or stewards operate and the possibility that service providers will act in self-interested ways (Knapp et al., 2011: 297).

Contracting out of employment services, which have operated in Australia for more than 10 years, is a notable case of governments purchasing services from community or not-for-profit agencies and private sector firms. After a brief outline of the employment services sector in Australia and how disability employment services fit into this broad context, this article uses the experience of disability employment services in Australia to assess the extent to which agency theory and stewardship theory capture the key factors that drive the behaviour of principals and agents/stewards. It then discusses co-production as a possible alternative.

Both agency theory and stewardship theory have a number of nuances, but these are not discussed at length here. Such an analysis involves complex issues that deserve a separate article. This article outlines relatively briefly the key features of the two theories. As stated above, its goal is specific: to assess the extent to which these two theories provide useful insights in understanding the actions of those involved in the provision of disability employment services in Australia. Co-production has a clear distinguishing feature: that organisational goals will not be achieved unless those being helped also make a conscious effort to increase their employability. It is this three-sided, rather than bilateral, relationship that is important.

Empirical data ¹ are used to analyse the impact of new contracting arrangements on DES providers and their clients. During the second half of 2010, one-on-one interviews or small focus groups were held with 124 staff and 73 job seekers from 27 not-for-profit or community service provider agencies in inner metropolitan, outer metropolitan and regional locations in every State and Territory in Australia except the Northern Territory. At the time of interviews, contracts had not been awarded to DES agencies through a competitive tender process, although this change was expected, and in 2012, 80% of DES providers were obliged to undertake a competitive tender process. The interviews and focus groups with service provider staff explored the changes that had taken place since the introduction of new contracting arrangements 6–9 months earlier, the impact of these changes on how the agency operated and the nature of the services the agency was able to provide for their clients. Job seekers were asked to describe their interaction with the service provider agency (or agencies if they had had experience with more than one agency), the type of assistance provided by the agency and whether they thought other forms of assistance would help them find a job. The interviews and focus groups were recorded and transcribed. A thematic analysis of the transcribed data revealed distinct commonalities across geographic locations, agency type (generalist or specialist) and level of staff (chief executive officer (CEO), manager or front-line worker). Responses from job seekers were more diverse, in part reflecting the diverse nature of the client group in relation to age, type of disability and length of engagement with disability employment services. Nevertheless, job seekers were clear about what it was they valued in their interaction with service provider agencies.

Employment services in Australia

In 1998, government provision of employment services was replaced by a combination of commercial and community or not-for-profit agencies, known as the Job Network. Job Network providers were selected through a competitive tender process and funded by an outcome-based funding model that required completion of services and achievement of outcomes before individual claims for payment were made, with future contracts dependent on an agency’s ranking relative to other agencies. From its inception, Job Network was an active labour market programme in that unemployed individuals were required to actively look for work and accept suitable work if offered (Nevile and Nevile, 2006). Service providers were required to report non-compliance to Centrelink, a central government agency that manages all social security payments and provides residual employment services such as referrals to Employment Service Providers. While Job Network was replaced in 2009 by a new generation of employment services, known as Job Services Australia (JSA), the compliance-centred nature of employment services persists (Fowkes, 2011).

The introduction of quasi-market arrangements for agencies whose work focuses on providing employment services for people with disability proceeded at a slower pace than for Job Network agencies. Many of the agencies providing disability employment services were established in the mid-1980s in order to provide job seekers with an intellectual disability (ID) with an alternative to supported employment in what was then known as ‘sheltered workshops’. Initially, agencies providing open employment services received block grant funding from the Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA), with policy responsibility transferred from FaHCSIA to the Department of Employment, Education and Workplace Relations (DEEWR) in October 2004.

The transfer of policy responsibility to DEEWR marked the beginning of a gradual shift in administrative and funding arrangements with disability employment moving closer to the administrative and funding model that regulates mainstream employment services. Up until 1 July 2006, what is now known as Disability Employment Services (DES) primarily provided assistance to voluntary job seekers who needed long-term support to maintain employment. Since 2006, DES agencies have experienced a decrease in the proportion of their traditional client base (voluntary ID clients) and an increase in the proportion of clients who have mutual obligations, that is, those who are required to look for work. By 2010, 65% of DES clients had mutual obligations, although this was not distributed evenly throughout the country, with agencies operating in metropolitan areas reporting a higher percentage of mutually obliged clients than agencies operating in regional or rural areas (Nevile and Lohmann, 2011: 25).

Agency theory

As noted earlier, agency theory assumes that goal conflict between principals and agents is likely to occur because agents have more information than principals and will exploit this situation to maximise their own utility. Principals therefore use a mix of incentives, sanctions, information systems and monitoring mechanisms to align the actions of the agents with the goals of the principal (Van Slyke, 2007: 162). Employment service providers in Australia face a mix of incentives and sanctions, with the majority of financial payments conditional upon the realisation of government goals (employment outcomes) sustained over certain time frames (13 weeks and 26 weeks). However, it is very hard, if not impossible, to create a mix of incentives and sanctions that eliminates self-interested behaviour. Under a system such as Australia’s, where payments are linked to the achievement of specific time-based outcomes, agents can maximise their financial return by focusing their attention on job seekers who are easier to place, a practice known as cherry picking or creaming. Creaming can also be combined with parking where providers fulfil minimum requirements in order to receive an initial payment and then make little or no effort to find employment for the more disadvantaged job seekers because their chance of success, and hence, the agency’s chance of receiving an outcome payment, is low (Bonvin, 2008: 372).

Risk selection (creaming and parking) has characterised, and continues to characterise, the employment services sector in Australia (e.g. Bredgaard and Larsen, 2008: 347; Fowkes, 2011: 8; Marston and McDonald, 2006: 8; Murray, 2006: 29–38; Productivity Commission, 2002: xxxviii; Struyven and Steurs, 2005: 219), with each change in the performance management framework creating new incentives for gaming or opportunistic behaviour. For example, in 2009, when the government changed the key performance indicators for highly disadvantaged job seekers from non-vocational outcomes to the employment-related outcomes that characterise the rest of JSA, a financial incentive was created for agencies to transfer highly disadvantaged job seekers to another agency. The fact that such transfers often occur just before review periods leads staff to speculate that this gaming behaviour is in fact occurring:

So we have got clients who are pushed out of other services … just before their review at 26 weeks or what have you. Other services recognise they are not going to get an outcome from them so they move them on. By that time, most of their payments have been used anyway so the other service gets no benefit but has to undo any damage that was done in the previous relationship. It is a real mess. (Employment Consultant, metropolitan agency)

Other staff go further, describing a shuffling of clients who have multiple barriers and are not job ready between mainstream JSA and DES agencies:

When I was a manager of a JSA provider, if we had these sort of clients, it was, ‘well they should be in DES, they’re not right for our program’, and we would do everything we could to get them into DES. Now I’m the DES provider, ‘well they’re not really right here either’. (Manager, metropolitan agency)

Agency theory assumes that all agents will act to maximise their utility. While the continued presence of risk selection indicates that this assumption holds true in some cases, empirical studies also reveal that many agents are motivated by concerns other than financial self-interest as demonstrated by the response of a CEO to the suggestion that in order to survive, the agency engages in opportunistic behaviour (job splitting):

We were advised … by an industry consultant … to go to the schools, get kids with learning difficulties and throw them at fast food outlets over Christmas. Then you will get your jobs up by the 31 December 2010 to feed into the 6 month outcomes that will be counted by 30 June 2011. And this is the strategy. Look for minimum outcome rate. For example, if someone has a 10 hour a week future capacity, stick them into a 10 hour a week job. If they want 25 hours, that is their problem, give them 10. You know, split jobs. I find this quite horrific. (CEO, metropolitan agency)

It is this type of scenario where professionals are encouraged or forced to act in ways they believe are inimical to the interests of the client that led Broadbent et al. (1996: 264) to question the application of agency theory as the design principle in the context of the ‘caring professions’, where control mechanisms imposed by principals clash with professional norms. In the case of disability employment services, conflict arises not through disagreement over policy objectives – like the government, service providers want their clients to find and sustain employment – but over how this objective is achieved:

We’ve got someone here that we knew, if she became physically active, that her pain would be better … and she would probably succeed in a placement, but we couldn’t justify it. Like there was no way DEEWR would ever accept it. ‘Why are you going to pay for her to get a membership to a swimming pool? That has got nothing to do with getting a job’. So we paid for a physio assessment and got the physio to recommend that water therapy would be beneficial … [but] we shouldn’t have to go through that. We should be able to identify that it’s needed and then do it. (Manager, regional agency)

DES staff clearly see themselves as professionals with the responsibility of acting in accordance with professional standards:

[W]e might see the perfect job for that client but it might be 12 hours a week [and their] benchmark is 30.We really have to think, ‘are we going to offer that job to that client or are we going to give it to someone else?’ … In one or two years we might be able to move [the client] to 30 hours, so we would get what the government is wanting us to get, but we would be allowed to be professional and do it in the timeframe. (Manager, regional agency)

All DES agencies are required to comply with National Disability Service (NDS) Standards and are audited by independent quality assurance auditors to ensure compliance. DES staff are happy to undergo this form of performance monitoring because they do not see any conflict between the values underlying the Disability Service Standards and the professional values that drive their work. What does frustrate staff is when the funding department prescribes how they will achieve employment outcomes without sufficient understanding of the nature of different types of disability: in other words, when staff see their professional expertise being undermined by those who have less professional knowledge and expertise. For example, the requirement to review and update a client’s Employment Pathway Plan each fortnight is seen as unnecessary. Staff explained that clients with an ID may be moving forward, ‘but very, very slowly’, and may take 18 months or 2 years to achieve their goal. For these clients, ‘what is the point of putting stuff in saying, “we are going to do this, or going to do that” [just] because it looks nice, it looks like you are helping?’ (Employment Consultant, specialist ID agency). Even with other clients, ‘a lot can happen in a fortnight, and nothing can happen’ (Employment Consultant, regional agency). If nothing has happened, reviewing an Employment Pathway Plan becomes another ‘tick the box’-type exercise, using up time that could be spent more productively.

It is the failure of agency theory to recognise mimetic isomorphism as a driver of behaviour on the part of agents that lies behind Frey’s (1993) argument that a mismatch between the motivations of agents and the type of incentives or sanctions imposed by principals reduces the effectiveness of those incentives or sanctions. For example, it has been argued that individuals willing to work in relatively low-wage sectors (such as the community sector) do so because of intrinsic motivations – they enjoy their work that provides feelings of personal worth or accomplishment (Alford and O’Flynn, 2012). Intrinsic motivation is fostered by (among other things) how much autonomy an individual is given in carrying out required tasks; external interventions, such as intensive monitoring, can have a negative effect on intrinsic motivation, causing the agent to decrease, not increase, work effort, (Frey and Jegen, 2001: 601), or leave the organisation (Davis et al., 1997: 40). In the disability employment sector, the negative effect of reductions in autonomy on intrinsic motivation is clearly evident in the loss of experienced staff, frustrated that they are no longer able to provide the sort of assistance they believe will best assist their clients:

What we are finding is that people who have been long-term working in the industry are leaving the industry. They are not moving between providers, they are going to work for the more welfare-based, socially holistic organisations that are still providing that type of service … This is unfortunate because they are not going to come back. (Senior manager, metropolitan agency)

Agency theory not only fails to recognise the lack of fit between financial incentives and what motivates service providers but also fails to take into account the motivations of clients, relying instead on the coercive power of sanctions. Under a compliance-centred regime such as Australia’s, job seekers are obliged to undertake certain activities, with service provider organisations acting as agents by reporting non-compliance that could result in financial penalties. As one job seeker explained, ‘… there is a power imbalance. You have to do what they say’. Other job seekers agreed, describing some front-line staff as ‘very forceful’:

Like originally I always wanted to be in admin … And my case manager basically told me, ‘you can’t do that, you have to do call centre’, and I kind of felt bullied into saying yes.

[Service X] would tell me, ‘we’ve got you a job, you’re turning up tomorrow or else we’ll take your payments away from you’.

For some clients, coercive power does engender compliance, but others resist, for example, by refusing to sign their Employment Pathway Plan or documentation stating hours worked, or by turning up to the agency and saying,

I am only here because Centrelink told me I have to. Don’t talk to me. I am just going to sit here for a while and then I will leave and you can’t do anything about it. (Employment Consultant, metropolitan agency)

Other clients employ covert resistance strategies, for example, turning up to an interview in bare feet, or ringing an employer about a vacancy and saying, ‘look I have a bad back and I can’t walk but I want to apply for this job’ (Manager, metropolitan agency). In some cases, coercive sanctions not only fail to engender compliance but also trigger, or exacerbate, anger and frustration, which manifests in aggressive, threatening behaviour towards DES staff:

A few years ago clients had a right to complete privacy and interviews were conducted in closed interview rooms. Yeah, I don’t think so now. You can’t be shutting doors now, it’s dangerous. (Staff focus group, metropolitan agency)

As the above discussion reveals, agency theory does not provide an appropriate way of conceptualising the delivery of complex products such as employment services. In such cases, it is impossible to devise controls that will eliminate opportunistic behaviour on the part of (some) agents and ensure compliance on the part of all clients. Furthermore, the most common control mechanisms used by principals – financial rewards and sanctions and intensive performance monitoring – tend to have a negative effect on those providing the service. Consequently, it is not surprising that stewardship theory, with its focus on intrinsic rather than extrinsic motivation, has been put forward as an alternative framework.

Stewardship theory

Under a stewardship framework, service providers are no longer agents who must be controlled by the principal, but stewards trusted to work co-operatively with the principal to maximise organisational objectives. As such, stewards should be empowered by the principal and their autonomy extended (Davis et al., 1997: 25; Knapp et al., 2011: 297), thereby eliminating the crowding out effects of intensive monitoring. Employment service providers have long been critical of inflexible, prescriptive contracts that limit the extent to which agency staff are able to use their professional judgement in determining how organisational objectives are to be achieved and would welcome a stewardship relationship that involved joint decision-making, information exchange and a principal that made an attempt to understand their needs (Nevile and Lohmann, 2011: 12–21).

While stewardship theory is a more accurate reflection of how the majority of service providers see their role, as with agency theory, its view of what motivates the majority of stewards is used as the basis for a normative model, which is applied to all stewards. In other words, stewardship theory assumes that all stewards are trustworthy and will not abuse the autonomy given to them by principals. A recent longitudinal study of Australia’s employment services sector revealed that the level of flexibility and autonomy given to service delivery agencies decreased between 1998 and 2008 (Considine et al., 2011: 817–819). Considine et al. (2011: 826) believe that one of the reasons for the increasingly intensive performance monitoring of service providers by the funding department was that when employment services were initially privatised, service providers used the flexibility they enjoyed under early contracts to increase their profits rather than pursue key public policy goals.

Advocates of both agency theory and stewardship theory acknowledge the need to transform low-trust situations, such as characterise mainstream employment services in Australia, into high-trust, co-operative behaviour, and empirical studies emphasise the importance of social bonds in enhancing compliance. For example, Braithwaite (1995: 252) argues that regulators can change the attitudes of regulatees opposed to the regulatory regime by nurturing their desire to be part of the regulatory community. More recently, vertical social re-categorisation has been suggested as one way of building trust and respect; it involves principals emphasising a common vision and shared goals in order to convert an ‘us vs them’ orientation into a ‘we’ orientation (Knapp et al., 2011: 303). Originally used to analyse the relationships between boards (principals) and top managers (agents) of private sector companies, vertical social re-categorisation does not translate well into the public sector where principals face a different set of accountability requirements. For example, in Australia, as in other countries with Westminster systems of government, funding departments tend to be risk averse because it is the minister who is expected to take responsibility for implementation failures, or perceived failures, and it is the job of government departments to implement programmes in ways that allow the minister to mount a plausible public defence of government administration (Mulgan, 2010: 12, 17). In other words, for principals operating in the public sector, performance management is primarily about managing political risk, and emphasising shared goals will not reduce their need for a performance management framework that focuses on how well the agent/steward is performing (evaluation) and on control.

In the employment services sector, evaluation indicators predominate (e.g. the proportion of job seekers placed in jobs 3 months after participating in job search training), with control achieved by linking funding to the achievement of particular outcomes. That is, inputs (financial resources) are dependent on outcomes. In theory, this should not be a problem if the outcome indicators are an accurate reflection of the full range of desired outcomes, or even of the intermediate steps or sequence of outcomes that will ultimately lead to the desired, final outcome. However, governments tend to focus on easily quantifiable indicators (for reasons of control) and usually do not elaborate sequences of outcomes, some of which may be hard to quantify. In the employment services context, payments are linked to clients finding a job or taking up education and training opportunities – outcomes that are easy to assess because the client is no longer in receipt of unemployment benefit. However, as those who work in this area know well, ‘hard’ outcomes – getting a job – are usually dependent on achieving ‘soft’ outcomes, such as improved self-confidence or self-esteem, particularly for disadvantaged job seekers. Yet, governments rarely reward achievement of these interim milestones:

I am thinking of having some tea groups myself. Social inclusion. There are people who need contact with other people … What price someone feeling they have made a friend when they are alone? Well, we won’t get paid for that, but if you want to get them employed, you want them to feel good in the world. (Employment Consultant, metropolitan agency)

Soft outcomes are more difficult to quantify and are often seen as more subjective than ‘hard’ outcomes. However, as noted earlier, a payment system focused on simple, narrowly defined outcomes provides incentives for creaming, while high initial payments (e.g. block grants) provide incentives for parking. To eradicate risk selection, a contract should reward providers ‘on the basis of their impact on the situation of each individual job seeker [with] the higher the impact, the higher the payment’ (Bruttel, 2004: 13). Therefore, contracts that reward interim milestones, such as improvements in soft skills, as well as rewarding employment outcomes and providing initial payments, are more likely to reduce the possibility of risk selection than appeals to shared goals.

Thus, just as agency theory fails to take into account the mix of objectives and motivations among agents (agency survival and/or growth as well as achieving outcomes for clients), stewardship theory fails to take into account the mix of objectives motivating principals, namely, achieving outcomes for clients and minimising political risk. Stewardship theory also fails to take into account the mix of objectives motivating clients. As noted earlier, all voluntary DES clients and the majority of mutually obliged clients share the common objective of securing and sustaining employment, but a small minority do not, and for this group of clients, non-coercive approaches such as vertical social re-categorisation are likely to be even less effective than when applied to funding department–service provider relationships.

Both agency theory and stewardship theory assume a vertical bilateral relationship, with service providers acting as agents for the funding department and managing the relationship with clients on behalf of the funding department. However, relationships are not bilateral because different parts of the bureaucracy maintain a separate relationship with clients from the one service providers manage on behalf of the funding department. For example, under the DEEWR contract, DES providers only get paid when their clients achieve their future benchmark (work capacity with intervention):

So really what the assessor is saying is that within two years if we give them assistance, they will get to 30 hours, but unless we put them in 30 hours now, we’re not getting any money for it. (Staff focus group, metropolitan agency)

Staff are very much aware that in many instances, capacity building can take months or years, and in the meantime, staff have to choose between pushing a client to take on more hours than they are presently capable of sustaining and getting an outcome payment, or working with the client to slowly increase capacity, but not getting paid for that work because the hours are less than the client’s designated future benchmark. Clients are confused when service providers respond to the extrinsic motivators built into the contract with DEEWR by focusing on future capacity, because Centrelink has told them that ‘you don’t have to work more than 15 hours, even though you’ve got a future capacity of 30 hours, you’ve got two years to get to your future capacity’ (Senior manager, metropolitan agency). At the same time, service providers assist clients to manage their relationship with Centrelink, an activity that is not part of their contract with DEEWR:

We have had problems for weeks now … trying to report income for that person with autism who is anxious about going down there [to Centrelink] and who doesn’t want to talk on the phone … So we need to assist him to do that. (Focus group of managers, regional agency)

In other words, relationships between the government, service providers and clients are triangular rather than bilateral. The existence of triangular relationships points to co-production as an alternative theoretical lens to either agency theory or stewardship theory.

Co-production

Co-production recognises that organisational goals will not be achieved unless clients contribute time and effort. In the employment services context, conscious effort is required on the part of the client to increase their job readiness (an intermediate outcome) as well as to secure and sustain employment (Alford, 2009: 99). The importance of clients taking an active role together with the service provider and the funding department in the production of outcomes is evident when employment outcomes for voluntary DES clients are compared to those who are obliged to participate. Apart from their eligibility for the Disability Support Pension (DSP), there is no clear difference between voluntary DES clients and those who are obliged to participate. ² However, what is clear is the difference in employment outcomes, with volunteers twice as likely to secure and sustain employment for 26 weeks than those who are obliged to participate (NDS, 2011: 5).

Furthermore, co-production recognises that clients, such as service providers and funding departments, are motivated by a mix of incentives, with intrinsic motivations being ‘at least as important as material ones in influencing clients’ willingness to co-produce’ (Alford, 2009: 131). Again, this more nuanced view is consistent with responses from DES clients who were not only appreciative of material incentives, such as assistance in getting a job or accessing training courses that interest them or are directly relevant to their career aspirations, but also valued non-material outcomes – being treated ‘as a person, not a number’, being treated with respect and listened to. Job seekers value agencies that take the time to ‘listen properly to what you require’ and do not just ‘put you in a job, any job’. As one job seeker puts it, ‘people should be treated as people – they have the right to choose’. The value placed on non-material rewards by DES clients is consistent with other Australian and international studies (see Nevile, 2008, 2009).

Like stewardship theory, co-production recognises that building self-esteem and activity directed towards achieving organisational goals depend on positive assistance, which is in line with clients’ intrinsic motivations (Alford, 2009: 133), but, unlike stewardship theory, co-production allows for a mix of incentives and sanctions, which ‘suggests how coercion might be properly situated in relation to voluntary exchange’ (Alford, 2009: 218). That is, coercion constitutes a deterrent for those clients who are resistant to the idea of participating in job search activities and acts as a ‘reserve backdrop’ for more willing clients, enabling service providers to focus on non-material rewards (Alford, 2009: 204).

Experienced DES staff understand the importance of non-material rewards, doing what they can to provide social activities for their clients, such as sausage sizzles or a Christmas party, or if there is no money for even these modest activities, staff do what they can to establish a relationship with each client as an individual while the client is in the office:

My clients respond to being treated with respect because this is one of the few places they’re actually going to get it. If I’m going to make someone believe that they can do something, that takes time and it also requires me listening to them and them knowing that. (Employment Consultant, metropolitan agency)

Co-production’s emphasis on social exchange, which includes non-material rewards such as respect or recognition of status as well as tangible rewards such as financial resources (Alford, 2002: 341), points to one way in which trust can be developed between risk-averse principals and service providers. Like clients, service providers value non-material rewards (autonomy and recognition of their professional competence) as well as material rewards necessary for organisational survival or growth. Sabel (2004) argues that the tension between the principal’s need to account for the expenditure of taxpayer’s money by minimising gaming behaviour on the part of (some) service providers and the desire of service providers for some level of autonomy in how organisational objectives are to be achieved disappear if the direction and substance of the exchange between purchasers and providers are reversed. Under a hierarchical, principal–agent model, accountability means reporting on and compliance with benchmarks, rules or standards imposed on the agent by the principal. The issue of accountability is not addressed directly in stewardship theory because it is assumed that the interests of the steward are aligned with the interests of the principal, and hence, stewards will always strive to realise organisational objectives rather than personal objectives. In what Sabel calls an experimentalist or pragmatic approach, provisional, initial goals are chosen and then revised in the light of more detailed, partial proposals, which arise from efforts to implement the initial goals. Because the ‘rules’ or benchmarks in this model are being continuously evaluated and changed if necessary, accountability equates to reason-giving, rather than compliance. That is, service delivery organisations are called upon to explain their use of the autonomy that they have been given in pursuing corrigible goals. Sabel’s approach can be seen as an extension of relational contracting, which is characterised by consensual or incremental decision-making (De Hoog, 1990: 325), but differs from relational contracting models in its emphasis on reason-giving as a form of accountability, rather than compliance with externally imposed or negotiated goals or benchmarks.

Under Sabel’s approach, monitoring is continuous and less concerned with outcome measures than with diagnostic information, that is, information that tells the service delivery agency and the funding agency what needs to be changed. Continuous improvement is the goal, and the response of the funding agency to information that suggests change is necessary is increased assistance to enhance the capacity of the service delivery organisation. Punishment (withdrawal of funding) only occurs after a service delivery organisation repeatedly fails to use the additional assistance provided by the funding agency to make the necessary changes (Sabel, 2004). This approach to monitoring provides a much better fit between the motivators used by those purchasing the service and the motivations of service providers whose intrinsic motivation is enhanced when they feel that the purpose of external intervention is to provide support rather than control their actions (Frey and Jegen, 2001: 594–595).

Sabel’s focus on diagnostic information is consistent with Behn’s (2003: 593) conclusion that outcomes are not necessarily the best measure of performance for all purposes. When the goal of performance measurement is learning and improving, Behn (2003: 593) recommends the use of disaggregated data that can reveal deviations from the expected (learning) and information about what is going on inside the organisation, which explains how changes in inputs, environment and operations lead to changes in outputs and outcomes (improving). Sabel’s approach is also consistent with a growing body of empirical and theoretical literature that highlights the importance of a new modern form of trust in generating and sharing new knowledge (Adler, 2001: 220). This modern, or reflective, form of trust is based on open dialogue among peers (Adler, 2001: 227), and as such is earned rather than assumed (Brown and Calnan, 2010: 20). Thus, reflective trust is not blind, it does not assume that professionals should not be held to account. However, it does assume that those who are being held to account are involved in the design of meaningful performance indicators (Broadbent et al., 1996: 280).

At first glance, an approach focused on learning and improving and built on reflective trust rather than control seems unlikely to appeal to risk-averse principals. However, Lynelle Briggs, former Public Service Commissioner and now CEO of Medicare Australia, recently noted that ‘achieving real reform in service delivery means thinking about issues and solutions in new ways’ (Briggs, 2010: 22). For Briggs, the core principles underlying service delivery reform will only be achieved if government departments put the service user first. While Briggs believes that success will be measured through the usual performance output measures, she does acknowledge that judgements about what constitutes success will be based on different sources of information, with the views and feedback from the community on service effectiveness and efficacy being vital in terms of future programme development (Briggs, 2010: 5). In the employment services context, this means that the funding department must move beyond simply seeking feedback from clients and service providers to using that feedback to develop and then adjust corrigible performance measures or benchmarks.

Conclusion

Governments contract out the provision of complex human services in the expectation that quasi-markets will deliver more efficient, flexible and innovative services. In the employment service sector, at least, these aspirations have proved elusive because regulations introduced to control gaming behaviour on the part of (some) service providers have tended to constrain rather than facilitate flexible, innovative service delivery, and the use of a competitive tendering model has encouraged an institutionalised lack of trust between the funding department and service providers. While governments are looking for responsive, innovative services, service providers are looking for more collaborative partnerships with government where their professional expertise is recognised and valued. Australia has had well over 10 years’ experience of a privatised employment services sector and as such is an important case for assessing the applicability of agency and stewardship theory to the delivery of complex human services. Based on the Australian case, this article argues that at best, agency theory and stewardship theory provide partial explanations of the relationships involved in the production of employment services. While elements of agency theory and stewardship theory are reflected in the empirical data, both agency theory and stewardship theory fail to account for the mix of factors driving the behaviour of service providers and principals.

Agency theory recognises powerful extrinsic motivators – the desire for financial rewards on the part of service providers – but does not recognise equally powerful intrinsic motivators – the desire to provide a service that conforms to professional standards or notions of ‘best practice’. Similarly, stewardship theory recognises the intrinsic motivators driving the behaviour of service providers but fails to recognise the particular accountability requirements of government departments and ignores the possibility that for some service providers, extrinsic motivators may prove stronger than intrinsic motivators. Both theories struggle to provide pathways that could reconcile the tension between the desire for flexible, innovative services and public sector accountability requirements. Furthermore, both agency theory and stewardship theory ignore the role played by clients in the production of employment outcomes, focusing instead on the bilateral relationship between the government funding department and service providers.

Given the limitations of both agency theory and stewardship theory when applied to the delivery of complex human services, this article argues that the concept of co-production is a more appropriate way of thinking about relationships between the government funding department, service providers and job seekers. Viewing job seekers or clients as active agents in the process of finding and sustaining employment suggests that engaging with job seekers as individuals and responding to individual needs and circumstances are likely to achieve better employment outcomes than the current ‘one-size-fits-all’ approach. Indeed, the government recognises the need for ‘individually tailored and comprehensive services’ (Australian Government, n.d.), but DES contracts remain prescriptive, substantially limiting the capacity of DES providers to deliver flexible, individualised services. With 60% of all DES job seekers not achieving any employment outcome (DEEWR, 2011: 5), a change in approach is clearly needed.