The Changing Nature of Contracting and Trust in Public-Private Partnerships: The Case of Victorian PPP Prisons

There has been increasing interest in the practices sustaining New Public Management (NPM) (Lapsley, 1999, 2009), with a range of research studies highlighting their similarities and differences across jurisdictions (see Olson et al., 1998; English and Skellern, 2005). One practice that is central to NPM involves government contracting with the private sector for the provision of infrastructure and various goods and services (Hood, 1995; Lapsley, 1999). A particular and significant instance of this involves the formation of public-private partnerships (PPPs), established by long-term contracting between public and private entities. In short, PPPs involve the use of private funds, risk-taking and management skills to provide public infrastructure and related services.

It is assumed that PPPs deliver infrastructure and services more effectively and efficiently than conventional public procurement processes (Spackman, 2002; Andon, 2007). Various PPPs have been created globally to provide hospital buildings and services, major infrastructure (such as roads, tunnels and bridges), schools, public housing, and prison accommodation and services (Harding, 1997). Extant research attests to the central role of contracting in the formation and operation of PPPs (Andrew, 2007; Andrew and Cahill, 2009), with recent studies outlining particular contracting practices (Broadbent et al., 2003, 2004; English and Walker, 2004). Nonetheless, it is acknowledged that much remains to be known about the detailed and changing nature of contracting in PPPs (Seal, 1999), as well as the ways in which contracting both informs and is informed by trust (Broadbent et al., 2003, 2004).

This article focuses on the changing nature of contracting and of trust in relation to PPP prisons in the Australian State of Victoria between 1995 and 2006.¹ With respect to Victoria, the Kennett Government (1992–99) surpassed other Australian states in its enthusiasm for NPM and the use of PPPs to outsource government services (English and Guthrie, 2003). Victoria's correctional PPPs account for 10.9% of the total number of Victorian PPPs, being worth $1,055 million and representing 7.7% of this state's PPP costs (English, 2006). Post-2000 and after the electoral defeat of the reformist Kennett Government, arguably a more tempered and less ideologically driven approach to PPP implementation became evident in Victoria (English, 2005).

The remainder of this work considers literature germane to our understanding of contracting and its connectedness to trust. This is followed by an examination of the context of, and the contracts used to commission, the PPP prisons. Overall, it is revealed that there has been marked contractual simplification in the post-2000 period, with the government's decision to eliminate private sector responsibility for the delivery of core correctional services, and a corresponding reliance being placed on the development of goodwill trust.

PPP CONTRACTING AND TRUST

The role of contracting in relation to hybrid, inter-organizational transactions, which characterize the formation and operation of PPPs (Baxter and Chua, 2003; Kuranmaki, 2004), has been the focus of a number of recent accounting research studies (Van der Meer-Kooistra and Vosselman, 2000[hereafter V&V]; Broadbent et al., 2003, 2004, 2008). Hybrid transactions are a contemporary manifestation of extended forms of contracting. This type of contracting aims to promote long-term inter-organizational collaboration while maintaining the autonomy of the contracting parties (Lorenzoni and Baden-Fuller, 1995), which in this case involves the state and a private prison operator. Saliently, an ‘ideology of trust’ (Mouritsen and Thrane, 2006, p. 243) is argued to be central to contracting in this context.

Borrowing from Williamson's research (1975, 1981, 1991, 2005), it is argued that inter-organizational forms of contracting are influenced by three important considerations—bounded rationality, opportunistic behaviour, and asset specificity. First, bounded rationality is important because it introduces uncertainty into contracting, sustaining a lack of ex ante‘presentiation’ or ability to anticipate completely future complexities and changes in a contractual transaction and its environment (Broadbent et al., 2003). As a result, PPP contracts are invariably ‘incomplete’. Such incompleteness requires the state and private prison operators to engage in ‘gap-filling’, a form of reflexive social monitoring (Broadbent et al., 2003) and improvisation, facilitating contractual implementation over the long-term timeframe of PPP agreements. Second, opportunistic behaviour recognizes the risk that self-interested and non-adaptive behaviours may prevail when one or both of the contracting parties fail to deliver specified outcomes when unforeseen events intervene. Third, asset specificity or the degree to which assets can be redeployed to alternative uses without sacrifice of productive value (Williamson, 1991, p. 281) increases the bilateral dependence of the contracting parties. Consequently, the continuing workability of PPP contracts becomes a matter of increasing importance in situations where highly specific assets, such as customized prisons facilities, are central to contractual performance. Lonsdale (2005) has indicated that the relatively high degree of asset specificity embedded in PPP contracts, as well as associated switching costs, may result in a form of ‘asymmetric lock-in’ (p. 72) between the state and the private provider. To promote the functional operation of PPP contracts, hybrid contractual transactions are presumed to be enacted by invoking a range of ‘elastic’ contracting mechanisms, as well as various ‘obligational market exchanges’ (Williamson, 1991, p. 271). The latter includes dispute resolution procedures, periodic competitive bidding/recontracting, and a material adverse effects regime in the case of the PPP prison contracts examined.

Given this, there is an emerging perspective that efficient long-term contracting relies on the emergence of a relationship between the contracting parties, which is characterized by flexibility and adaptability (Broadbent et al., 2003, 2004). This view recognizes ‘the limit of economic rationality to explain contracting, and places transaction costs in the context of their social relations’ (Broadbent et al., 2003, p. 177). Campbell and Harris (1993) present evidence indicating how, in practice, contracting parties work to preserve a relationship, despite difficulties in adhering strictly to contractual obligations. This has led some researchers to conclude that contracts ‘are little used in controlling the ongoing relationship’ (Broadbent et al., 2003, p. 178), being invoked mainly in situations requiring arbitration or the interpretation of a court. In comparison, however, other research studies emphasize the range of emergent uses that contracts may also assume in facilitating the functioning of hybrid transactions. For example, both Dekker (2004) and Tomkins (2001) highlight the role of contracting in facilitating inter-organizational goal-setting, planning and coordination.

The functionality of adaptive and flexible behaviours in facilitating the realization of objectives in long-term contracting has led researchers and practitioners to argue that the operation of hybrid relationships requires a degree of systems or impersonal trust (V&V, 2000; Chenhall and Langfield-Smith, 2003; Busco et al., 2006).² PPP contracts are no exception to this argument, with trust being considered essential to efficient and effective contracting (Spackman, 2002; Broadbent et al., 2003, 2004). Given its capacity for uncertainty reduction, trust is construed as being central to the management of risk (Langfield-Smith and Smith, 2003), with risk management residing at the heart of PPP contracting. Following Broadbent et al., trust is argued to moderate the effects of uncertainty embedded in: first, the transaction duration, which may be up to sixty years for some PPPs; second, the transaction environment, which may involve substantive shifts in the political and policy environments over the lifetime of a PPP; and third, between the transacting parties, given that the state and private providers bring quite different capabilities and expectations to a PPP agreement. Nonetheless, as V&V and others have found, while the operation of some hybrid transactions involves a heavy reliance on trust, this can be tempered by contractual provisions introducing both bureaucratic-like rules to mitigate problems associated with incomplete contracting, such as the detailed specifications of tasks and required service delivery outcomes in PPP contracts, and/or market-like mechanisms, which include the recontracting windows that operate at various stages in PPP relationships.

Further, it is argued that different constructions of trust sustain efficient long-term inter-organizational contracting (see V&V, 2000; Broadbent et al., 2003; Langfield-Smith and Smith, 2003). While many different typologies of trust have been proposed (see Mitzal, 1996), the notion that contractual, competence and goodwill forms of trust facilitate contracting is most common in germane PPP research (Broadbent et al., 2003, 2004). Contractual trust arises from written guarantees (Chenhall and Langfield-Smith, 2003) and sustains expectations that the nature of an exchange will be as anticipated, that is, a private provider of infrastructure and/or services will work to fulfil the terms of a PPP contract and the state will likewise meet its responsibilities. Competence trust denotes a belief that a partner possesses the relevant technical skills and managerial expertise (Broadbent et al., 2003, p. 179), with the private operator, in this case, having the capabilities to build and/or operate a prison on behalf of the state. An inverse relationship has been argued between the level of competence trust and the need for ex post inspection by the purchaser of goods and services (V&V, 2000; Dekker, 2004). Goodwill trust is used to denote expectations of dependability in terms of the contracting parties' emergent behaviours (Langfield-Smith and Smith, 2003). This is reflected in adaptive practices that move the contracting parties beyond their original promises in order to ensure the ongoing viability of a relationship. Goodwill trust especially is argued to develop over time and as a result of repeated interactions, which engender mutual learning and adaptation between the contracting parties (Tomkins, 2001; Langfield-Smith and Smith, 2003).

Goodwill trust, and its manifestation in terms of a relationally adaptive commitment to the long-term functioning of an agreement, is seen as promoting goal alignment between the contracting parties (Stephen and Coote, 2007). Goodwill trust is also seen as encouraging and enabling innovations to achieve this, in relation to shifting aspirations and emerging conditions over time (Cooper and Slagmulder, 2004). Broadbent et al. (2003) argue it is relational contracting that allows change to ‘be dealt with’ in PPP contracts (p. 178). In their study of the Private Finance Initiative (PFI), they outlined that the U.K. government implemented a number of institutional-level initiatives to build goodwill trust by assuring private contractors of the government's long-term commitment to the PFI. In that instance, it involved reducing private sector exposure to default risk and formally designating a role for performance measurement in the calculation of incentives and penalties, including associated abatement and default regimes.

While there is evidence of relational expectations informing the execution of some PPPs (Seal, 1999; Broadbent et al., 2003), there are also contrary claims of private suppliers barely meeting envisaged contractual service levels. It is pertinent that examples of such claims have been made with respect to contracting in the penal sector (Cooper and Taylor, 2004; Andrew, 2007), implying a lack of not only goodwill trust but also of competence and possibly contractual trust in this context. Furthermore, Broadbent et al. (2003) outline that disagreements regarding the availability and performance of complex assets, such as hospitals and prisons, may also inhibit the emergence and development of trust between the state and private contractors—in part because of the limitations of accounting performance measures to capture contractual expectations in complete, simple and controllable terms. Additionally, Broadbent et al. argue that there may be a contradiction between the achievement of value for money (VFM) and the need to develop goodwill trust in order to manage a long-term PPP relationship in an essentially bureaucratic (public-sector) system. For example, in the particular NHS Trust that Broadbent et al. observed, there was no explicit attempt to evaluate whether the required risk transfers and VFM had been achieved. Instead, accounting performance measures were used as an imperfect and indirect means for managing the accomplishment of this key contractual objective.

More fundamentally, however, Froud (2003) has contested the veracity of the constitution of goodwill trust in PPPs, questioning the possibility of flexible and adaptive forms of such contracting. Instead, she argues that PPP contracts cannot deliver PPP policy objectives or manage and control future uncertainties, stating, ‘the PFI contract reduces the ability of the public sector to deal with uncertainty, by locking the state into contracts typically over 20 to 30 years and reducing the flexibility to respond to a dynamic environment’ (p. 582). Froud denies the development and utilization of goodwill trust to manage uncertainty in the transaction environment of PPPs because she assumes that the necessary and repeated interactions between the state and private providers of assets and services do not occur. This proposition, however, is difficult to support in relation to the PPP prison contracts examined in this research in which a variety of mechanisms require ongoing communication and engagement.

Of relevance also to an understanding of the potential inter-relationships between PPP contracting and trust is research outlining possible connections between trust and the role of accounting information and controls embedded in the governance of inter-organizational relationships. This research is summarized by Mouritsen and Thrane (2006). First, Mouritsen and Thrane contend that some accounting research, such as Tomkins (2001), portrays the constitutive capability of accounting controls: accounting controls build trust in inter-organizational alliances such as PPPs. It is maintained that the act of sharing information between contractual partners for management and control facilitates alignment through the development of a shared framework, which is amenable to the formation of stable, realistic and common expectations required for trusting behaviours. A second strand of research identified by Mouritsen and Thrane is represented by the work of Dekker (2004) and V&V (2000). It portrays a substitutive rather than complementary role between trust and accounting controls: trust between contracting parties replaces a need for accounting controls that monitor behaviour and outcomes. Finally, Mouritsen and Thrane depict Cooper and Slagmulder's (2004) research as providing a further and plausible possibility in which trust acts as a platform for the development of ‘advanced forms of accounting’ (Mouritsen and Thrane, 2006, p. 243) and control. On reflection, Mouritsen and Thrane remark that trust is ‘more likely an aspect of evaluation, a statement about how a relationship can be evaluated, and hence becomes a marker about the desirability of certain practices, including accounting practices’ (p. 243).

Overall, the above suggests that relational contracting is necessarily connected to the accomplishment of PPP agreements. Moreover, the emergence of relational contracting in this context turns on the realization of contractual, competence and especially goodwill forms of trust between the state and private providers of various forms of assets and services. Yet it is acknowledged that much remains to be learned about detailed contracting practices and the ways in which these impact on, and are impacted by, various constructions of trust, especially in relation to PPPs. It is also the general conclusion of Broadbent et al. (2008) that such understandings of the practices informing the workings of PPPs in different times and places need to be established empirically, with Lapsley (2001) further encouraging longitudinal case studies as a means of achieving this. It is therefore our intention to engage with these thrusts from the extant literature. In particular, we narrate a case study investigating the ways in which contractual, competence and goodwill trust are connected to contracting practices that encourage the development of relational contracting in the context of Victorian PPP prisons agreements.

STUDYING THE VICTORIAN PPP PRISON CONTRACTS

The State of Victoria has commissioned five PPP prisons.³ Three—Deer Park Women's Metropolitan Prison, which is now known as Dame Phyllis Frost Correctional Centre (hereafter Deer Park), Fulham Rural Men's Prison (hereafter Fulham), and Port Phillip Men's Metropolitan Prison (hereafter Port Phillip)—were commissioned by the Kennett Government in the pre-2000 period. The other two—Marngoneet Correctional Centre and Melbourne Remand Centre (hereafter Lara and Ravenhall respectively)—were commissioned by the succeeding Bracks Government in the post-2000 period.

As confirmed by English and Walker (2004), the State of Victoria is a party to two primary contracts for each prison agreement. One establishes the Prison Services Agreement (PSA) in the case of the pre-2000 contracts or a Facilities Services Agreement (FSA) in the case of the post-2000 contracts. The other contains the Tripartite Deed, the primary financing document, recognizing the state as the financial guarantor of these prisons.

The contracts for the PSAs/FSAs constitute the primary sources of data used in this research.⁴ As these documents are durable and unaffected by the passing of time, this arguably mitigates some of the reservations that may surround the use of a retrospective case study (Yin, 1989). It is noted, however, that the two post-2000 FSAs omit some of the contractual detail made available in relation to the pre-2000 PSAs. A lack of comparable contractual disclosure is acknowledged as a limitation of this research.

Various secondary documentary sources are also utilized, including pertinent acts of parliament (Corrections Act 1986), Victorian Department of Justice (VDJ) policies (VDJ, 1994, 1997), independent investigations into Victoria's PPP prisons (Kirby et al., 2000), consulting reports (KPMG, 2000), reports of the Victorian Correctional Service's Commissioner (Office of Correctional Services Commissioner [OCSC], 2000), Victorian Department of Treasury and Finance guidelines and reports (PV, 2003, 2005; VDTF 1991, 1994, 1996), audit commission reports commissioned by the Victorian Government (VCA, 1993), and reports by the Victorian Auditor-General and the Australian National Audit Office (VAGO, 1999, 2001, 2003, 2007; ANAO, 2004).

The contracts are analysed to characterize constructions of contractual, competence and goodwill trust and how these change over time. While it is acknowledged that these three constructions of trust are intertwined in practice, for the purposes of this research they have been delineated and mobilized empirically in the following ways. Contractual trust is characterized in terms of written expectations (Chenhall and Langfield-Smith, 2003), which establish the objectives of PPP prisons and the responsibilities of the contracting parties. Competence trust is constructed by contractual provisions highlighting the scope and level of expertise expected from the private sector partners (Langfield-Smith and Smith, 2003). Goodwill trust is evidenced by explicit contractual provisions outlining the ways in which uncertainty is to be managed by the contracting parties, as well as emergent practices sustaining the discretion of the contactors in relation to the achievement of desired contractual objectives (Langfield-Smith and Smith, 2003). This analysis is accomplished by identifying the intention of each contract and how this is realized through explicit written guarantees and unwritten expectations, which are informed by the policy context of the contracts. By using examples from the pre- and post-2000 contracts and supported by information derived from secondary sources, the nature of changing constructions of trust embedded in the contracting practices is explicated and then discussed.

CONTEXTUALIZING PPP PRISON CONTRACTING AND TRUST

An understanding of the changing nature of contracting and trust in these PPP prison contracts is necessarily informed by an appreciation of prevailing social, political and policy environments, including shifts in the ideological commitment to PPPs by the Victorian government (Scott, 1990). It is assumed that observed changes in contracting are reflective of deeper and wider changes in government objectives concerning the management of relationships between the private and public sectors. Given this, the analysis is guided by a contextual perspective that attempts to understand how and why the phenomena of interest have changed over the period investigated, as well as the significance of the observed changes (Peraekylae, 2005). As such, this section outlines how the accomplishment of contractual, competence and goodwill trust are envisaged to support the achievement of outcomes in relevant policies that were embedded in the pre- and post-2000 contracts.

The involvement of the private sector in service delivery, and the magnitude and ambition of the associated New Prisons Project reforms promulgated by the Victorian Department of Justice (VDJ) in 1993 (VDJ, 1994, 1997), are predicated on an unquestioning belief in the superiority of private sector solutions (Laughlin and Pallot, 1992). Nonetheless, there are significant differences in the transaction environment and the scope of private sector responsibility for prisoner management between the pre- and post-2000 PPPs. Table 1 provides an overview of the transaction characteristics of the three pre-2000 PSAs and the two post-2000 FSAs. It reveals differences between the Kennett and Bracks Governments' approach to using private contractors to provide prison infrastructure and associated services. Notably, Table 1 provides indicators that contradictory elements in the transaction environment, such as the risk transfer associated with an off-balance sheet accounting treatment and the levels of trust required to support the magnitude of contractor responsibility for the management and rehabilitation of prisoners in the pre-2000 environment, may undermine the achievement of contractual objectives. Furthermore, different transacting parties and differences in the profile and number of prisoners in the pre-2000 arrangements prospectively complicate the VDJ's management of these contracts. By contrast, the post-2000 transaction environment ostensibly poses significantly fewer risks for the VDJ, largely due to the more limited contractor role and lack of risk transfer associated with the on balance sheet accounting treatment.

As outlined in Table 1 and more particularly with respect to the pre-2000 contracts, the contractors have sole carriage of the management of all core and non-core correctional services at the prison facilities, albeit in relation to prisons of varying size and prison profile. To this end, the passage of amendments to the Corrections Act (1986) and the promulgation of infrastructure investment guidelines and policy by the Victorian Department of Treasury and Finance (VDTF, 1991) enabled both the commissioning of private prisons, and the establishment of a systemic framework (Broadbent et al., 2003) within which the VDJ (1994, 1997) intended to comprehensively reform the housing, management, treatment, education and rehabilitation of prisoners. More particularly, and contrary to the PPP literature (Froud, 2003; Andrew, 2009), the private sector was entrusted with leading the modernization of the prison sector (see Lapsley, 2009), with the pre-2000 prisons being intended to spearhead the implementation of the outcomes-based New Prisons Project reforms to correctional services delivery.

The pre-2000 contracts reflect a new regulatory environment for the Victorian corrections industry, overseen by the purposefully created OCSC. For the first time, performance-based service-delivery outcomes were created and linked to payments for correctional and facility services throughout the prison sector. The intention was to create a commercial incentive for high quality services, with public prisons also to be funded and governed on the basis of similar contracts. This was supported by the introduction of comprehensive contractual safeguards disciplining poor performance and ensuring the promotion of the Victorian public's interests. Also, the contracts established the basis for the anticipated achievement of reduction in costs in service delivery by ‘the transfer to the private sector to the maximum extent possible, the risks of ownership of buildings and facilities and the risks associated with service delivery’ (VDJ, 1994, p. 11).

Arguably, this blanket risk transfer is associated with high project risk for the contracting parties. This may have hindered the realization of goodwill trust associated with the pre-2000 PSAs, especially as the contracting parties worked to mitigate uncertainties associated with the achievement of qualitative outcomes relating to the housing and treatment of prisoners enshrined in the Prison Management Specifications (PMS) included in the pre-2000 project briefs of the VDJ (VDJ, 1997). The political and social context of the contracts was central to this failure because it dictated the specifics of the contract, which, in turn, created unsustainable expectations of the relationships in practice.

By 2000, disappointing outcomes in other pre-2000 PPP projects and a changed political environment (signalled by the election of the Bracks Government and the establishment of Partnerships Victoria in VDTF) meant a different approach to PPPs was adopted in Victoria. Still committed to PPPs, the Victorian government significantly curtailed the role of contractors in the delivery of public services (PV, 2005) by ensuring that the correctional services are managed by public employees and not by private contractors, thereby significantly reducing project risk and its potentially adverse impact on the formation of goodwill trust. However, these contracts introduced a new risk—interface risk—due to the government's reliance on the contractor to provide facilities and security services in a manner that supports (rather than thwarts) the VDJ's management of prisoners.

A significant change in the post-2000 contracts was that ownership of the post-2000 facilities resides with the government, indicating that these arrangements are recognized in the state's financial statements. This confirms that blanket risk transfer is not an overwhelming objective in the post-2000 period (PV, 2005). The difference in accounting treatments (see Table 1) is also indicative of the government's growing appreciation (as evidenced in its post-2000 PPP policies [English, 2005]) of the important role played by goodwill trust in effective long-term contracting. The reduction in project risk resulting from the curtailment of direct contractor involvement in the delivery of core services is associated with the post-2000 projects. This signals the government's newly-found appreciation that contracting practices aimed at transferring high levels of project risk to the private sector can undermine the achievement of successful long-term contracting outcomes in PPPs.

The following section outlines in more detail the ways in which the contracts establishing the pre- and post-2000 PPP prisons in Victoria inform and are informed by changing constructions of trust, shaping the relationship between the state and the private prison providers.

THE CHANGING NATURE OF CONTRACTING AND TRUST

This section outlines the ways in which the State of Victoria crafted changed contracting practices in relation to pre- and post-2000 PPP prisons, reflecting the abovementioned shifts in the policy context of these agreements. In particular, the following subsection outlines how the pre-2000 contracts frame explicit expectations mediating the relationships between the state and private prison providers, informing the substantive nature of contractual, competence and goodwill trust. This is then followed by a similar consideration of the post-2000 PPP prison contracts.

THE EFFECTS OF CHANGING CONTRACTING PRACTICES AND TRUST

The above identifies five main changes in substantive contracting practices between the pre-2000 PSAs and the post-2000 FSAs. First, the objectives of the pre-2000 contracts are broader, with the private sector being responsible for the design, construction and financing of the prison facility and all aspects of prison services. The private contractor is responsible only for the commissioning of the facility and the provision of non-core prison services under the post-2000 FSAs. Second, the pre-2000 PSAs assign risks to the private sector, including risks associated with changes in government policy and prisoner profile, numbers and mix.¹³ The post-2000 FSAs, in comparison, aim to assign risks to the party best able to cope with them, with the material adverse effects regime and gain-sharing provisions being important mechanisms for achieving risk management and the development of goodwill trust. Third, the post-2000 contracts, unlike the earlier PSAs, involve self-monitoring and exceptions-based reporting processes. Fourth, the post-2000 FSAs introduce compensation by the state for changes in the scope of services or performance standards, as well as an abatement regime that monetarizes the impacts of underperformance. Fifth, and finally, the post-2000 contractual provisions provide a greater emphasis on collaboration, with clauses instituting formal roles to manage the interface between the state and the private sector and to promote greater interaction and information sharing. These post-2000 changes are also significant in that they overtly provide for the role of goodwill trust in the ongoing processes of contractual execution.

These changes have had significant effects in terms of the ways that trust is constructed by and between the pre-2000 PSAs and post-2000 FSAs. Changed contractual objectives have redefined the contractual and competence trust informing the operation of the PPP prisons. There has been a significant reduction in the scope of the contractual and competence trust expected of the private providers by the state in the post-2000 contracts. These have been limited to the construction of prison facilities and the performance of non-core activities only. The changed contractual practices of the post-2000 FSAs also indicate that the state has been prepared to introduce new contractual provisions to foster the development of goodwill trust, building on the bedrock of contractual and competence trust necessary to enable the commissioning of a prison facility. Whereas the pre-2000 contracting practices have tended to delimit and hinder the emergence of sustained goodwill trust from the private sector, because of the blanket risk allocation procedures adopted and the nature of the SDOs used in performance monitoring and rewarding, the post-2000 FSAs arguably anticipate the development of goodwill trust to enable their functioning as intended by the contractual objectives and informing policy context. The emphasis placed on the ongoing interface between the government and the private sector in the face of uncertainty, as well as the importance of self-monitoring, regular communication, and negotiated outcomes, are predicated on the emergence of goodwill trust (Williamson, 1991; Campbell and Harris, 1993; Broadbent et al., 2003; Stephen and Coote, 2007). Accordingly, the Victorian Auditor-General has acknowledged explicitly that the accomplishment of the post-2000 contractual objectives is premised on the performance of relational contracting:

However, it should be noted that government inexperience (English, 2005) and the contradictory objectives of VDTF and of VDJ with respect to the pre-2000 PPP prisons have hindered the development of relational contracting. A primary purpose of VDTF was to manage the macro-economic aspects of infrastructure provision by ensuring that the assets and associated liabilities were not recognized in government financial statements. In keeping with this macro-economic objective, VDTF was also concerned to find cost-effective solutions at the micro level (Broadbent and Laughlin, 1999), resulting in a predisposition to accept the lowest bid (VAGO, 2001). However, the auditor-general has noted that ‘it is not surprising’ (VAGO, 2001, pp. 221–2) that cost pressures affected the standard of service offered by the prison contractors, impeding the contrariwise objectives of the VDJ (encompassed in New Prisons Project) seeking improvements in the management, housing, treatment and rehabilitation of prisoners. This objective requires competent management and sufficient resourcing to enable flexibility to achieve the necessary changes in prisoner behaviour associated with reducing re-offending in complex and changing policy and social environments (Kirby et al., 2000).¹⁴

Interestingly, there have been different outcomes and performance with respect to the various relationships between the state and private prison providers, despite the commonality of the contractual provisions and the policy environment informing each of the prison operations. This has been particularly acute in relation to the pre-2000 contracts wherein similar contracts (Zhang, 2004) have produced quite different outcomes. A lack of contractual, competence and goodwill trust has been associated with the management problems leading to the Deer Park failure, and this has been partially explained on the grounds of poor partner selection and development (VAGO, 2001).¹⁵ At Port Phillip the contractor has struggled to achieve required performance standards (VAGO, 1999; Kirby et al., 2000), with the relationship being fraught at times. However, the department has acknowledged that the profile of Port Phillip makes it a very difficult prison to manage. This was compounded in the early years of the agreement by the negative impact on contractor performance of sub-optimal decision making by VDJ (Kirby et al., 2000) and unanticipated changes in the transaction environment, such as prisoner numbers, profile and mix. Nevertheless, as the 2006 recontracting process reveals, despite initial problems, there is now consensual agreement about the management of Port Phillip that better reflects the department's original New Prisons Project objectives, which are imperfectly captured in the 1995 contract. In the less complex environment characteristic of Fulham, the contractor has continued to demonstrate its ability to adapt and change to unanticipated circumstances (Kirby et al., 2000; Graffam et al., 2005; VAGO, 2007).

This suggests that uncertainties arising from the transaction environment and the transacting parties (Williamson, 1981, 1991, 2005; V&V, 2000) are more important to contracting outcomes than is currently acknowledged in the PPP literature. Indeed, these factors may combine to militate against goodwill trust development, despite the intentions of the contracting parties and the presence of contractual and competence trust. The transaction environments described in other private-sector (see V&V, 2000; Langfield-Smith and Smith, 2003; Mouritsen and Thrane, 2006) and PFI (Broadbent et al., 2003, 2004, 2008) case studies are less complex than that informing the pre-2000 prisons outlined in this research. The very different objectives and accountabilities of the transacting parties in the Victorian arrangements also add to the political and performance risks associated with the early PPP prisons (Kirby et al., 2000)—and help to explain the Victorian government's move to reduce these risks by curtailing the role of contractors in the post-2000 agreements. Differences in the transaction environment and transacting parties may also explain why the contracts continue to play a prominent role in ongoing interactions with the state. Contrary to assertions in the literature that contracts ‘are little used in controlling the ongoing relationship’ (Broadbent et al., 2003, p. 178), discussions with officials from VDJ and operators suggest that the contracts are referred to on a daily basis in ongoing interactions between the contracting parties. For those PPP prisons in which private contractors are delivering the gamut of core and non-core services, the contract has become the touchstone from which trust develops (or fails to develop), mediating repeated interactions in the interpretation and implementation of its provisions in an uncertain and shifting environment.

The prison contracts investigated also raise questions regarding how accounting practices relating to monitoring and rewarding mediate envisaged contractual, competence and goodwill forms of trust sought in this highly uncertain and complex transaction context. Arguably the contracting practices examined here present a more intertwined and multiplex relationship between accounting and trust than has been found in extant research (Mouritsen and Thrane, 2006). Consistent with Tomkins' (2001) research, the contracts mobilize accounting control practices to constitute trust, with the post-2000 self-reporting and exception reporting processes being evidence of this. In addition, there is also evidence of trust-based practices acting as substitutes for explicit accounting control processes. It is acknowledged that the SDOs embedded in the prison contracts are partial and incomplete, particularly when these are considered with respect to the intentions of the contractual objectives and the uncertain and shifting policy environment in which these PPPs are situated. However, the ongoing development of some prison programs under the auspices of the pre-2000 prison agreements (Kirby et al., 2000; Graffam et al., 2005) attests to the realization of goodwill trust that reflects the spirit of the contracts, rather than a narrow achievement of performance outcomes. Based on this, the prison agreements have fostered both substitutive and complementary roles between accounting controls and trust, suggesting further research is needed to understand the interpenetration of contracting and trust by drawing on real-time case longitudinal studies of contracting in this sphere of practice.

CONCLUSION

This article has documented the changing nature of contracting and trust that characterizes the pre- and post-2000 PPP prison contracts executed in the Australian State of Victoria. Moreover, it has done so distinctively, drawing on both unique primary documentary sources and an appreciation of the shifting political and policy contexts of the contracts examined in this research. The analysis has highlighted changing contractual practices with respect to the objectives of these PPPs, risk management, performance measurement and management, the structuring of payments and incentives, and the emphasis placed on collaboration and communication. In short, the pre-2000 contracts aimed to devolve the provision of all prison facilities and services, in conjunction with their attendant and emergent risks, to the private sector. These contracts and their implementation anticipated broad ranging forms of contractual and competence trust on behalf of the private sector. This has not always been borne out in practice because of the inherent tensions in the transaction environment, informed by a blanket approach to risk transfer and an off-balance sheet accounting treatment of these PPPs, impinging on the development of goodwill trust.

In comparison, the post-2000 contracts are much narrower in scope, returning the provision of key prison services to the public sector. The workability of these latter contracts is premised on a more collaborative relationship being forged between the state and private prison contractors. This is being reflected in revised approaches to risk sharing, the use of incentives, and a need for flexible and adaptive behaviours to realize the overarching spirit of these agreements. The accomplishment of goodwill trust acts as the cornerstone of the post-2000 PPP prison agreements, with the functioning of these agreements being underscored by a less complex and lower risk transaction environment.

More generally, the findings lead us to reflect—even if only tentatively and with a view to engendering future and more extensive researcher engagement with this particular field—on the role of contracting in relation to PPPs, both in penal and other sites of public sector activity. This study has implications, for instance, as to how PPP lifecycles are framed analytically. Our investigation into the changing nature of contracting in these specific Victorian PPPs indicates that contracting is a persistent and recursive practice that cannot be discretely delineated from the so-called operating phase of a PPP's lifecycle. In addition to the periods of recontracting required by the PPP agreements, contracting is shown to be an ongoing and organic process informing the accomplishment of PPPs, as both parties continue to re-engage with instances of poorly specified or omitted written intent, particularly in relation to desired qualitative outcomes and the changing transaction environment embedded in potentially varying PPP steering mechanisms and shifting community standards with respect to the treatment and rehabilitation of prisoners.

Against this backdrop, the contracts for the five PPP prisons in Victoria emerged as devices with the possibility to deliver more than discrete economic outcomes, such as value for money, risk allocations between the public and private spheres, and potential protection against the threats of opportunism. These contracts arguably transformed the social also (see Latour, 2005). The contracts in question mediate processes of sense-making, dialogue and learning about the capabilities of the contracting parties, the boundaries between the ‘public’ and the ‘private’, the day-to-day accomplishment of prison services, and the envisaged standards and outcomes sought from these agreements by various stakeholders. As such, this research suggests that PPP contracts act as important ‘funnelling’ and coupling devices in relation to the networks of agents, policies, practices and interests that form and sustain PPPs in action. It is imperative that we better understand these networks, and the roles that contracts assume within them, both shaping and shaped by the institution of particular PPPs.

Indeed, the importance of achieving an understanding of the relational context of PPPs was indicated in the conclusions drawn from this research. It was contended that the importance of differential contractual transaction environments and transaction party profiles are central to an appreciation of the performance of the PPP prisons in the State of Victoria. Similar contracts have led to very different contractual outcomes over the period investigated. Contractual refinement and development, while important, are insufficient to steer such relationships towards the achievement of their espoused objectives. Such contracts must be implemented in tandem with a cognate understanding of the contextual factors and local practices, including trusting behaviours, which mediate the situated accomplishment of a PPP agreement. Researchers are encouraged to study how these situated practices are mobilized in the context of the network of relational ties that constitute the functioning of a PPP.

By way of final comment, it is contended that studying the enactment of contracting and trust in PPPs is an area of practice that is ripe for more rigorous and varied theorization, as well as more granular accounts of the networks of practices enabling their realization.