Beyond ‘one size fits all’: how local conditions shape PPP-enabling field development

Abstract

The use of public–private partnerships (PPPs) for infrastructure development has received significant scholarly attention of late, but there remains a need for more work at the programme level. Specifically, there is a need for work that recognizes the way that PPP programmes are implemented differently in different regions, thereby progressing beyond an effectively ‘one size fits all’ view of PPP programmes. In response, this paper offers a comparative analysis of the historical development trajectories of three contemporary PPP programmes: in British Columbia (BC) (Canada), Victoria (Australia) and South Africa. We begin by recognizing the role played by the UK's private finance initiative as a programme model, and then show how this model was adapted and modified in each of our cases, leading to very different field structures. The study uses a grounded theory building approach and draws heavily on theories of institutional change and structuration. There are two main contributions from this study: (1) we draw attention to the need for a context-specific approach to explain and predict PPP field development and (2) we develop arguments relating to institutional systems and processes that can guide future studies of these and similar fields.

Keywords:

• Case study
• infrastructure
• institutional change
• organizational field
• programme level
• public–private partnerships

Introduction

The use of public–private partnerships (PPPs) for infrastructure development has received significant scholarly attention of late (Gil and Beckman, 2008; Kwak et al., 2009). The PPP concept is still somewhat contested (Hodge and Greve, 2005; Weihe, 2008), but the use of the term for infrastructure developments in the vein of the UK's private finance initiative (PFI) have dominated policy rhetoric internationally (Weihe 2008, p. 10). Previous research has primarily focused on various elements of PPP project development, including types and bearers of risk (Li et al., 2005a), value for money (Akintoye et al., 2003; Heald, 2003; Nisar, 2007), critical success factors (Li et al., 2005b; Zhang, 2005), project finance (Yescombe, 2007; Pretorius et al., 2008), project governance (Klijn and Teisman, 2000; Bult-Spiering and Dewulf, 2006) and the theoretical underpinnings of the PPP model (Savas, 2000; Bovaird, 2004; Hodge and Greve, 2005; Yescombe, 2007; Hellowell and Pollock, 2010).

PPP projects are, however, invariably situated within larger PPP programmes, i.e. more or less coordinated collections of PPP projects undertaken in a given politically sovereign area. Research that looks at this collective level has been much more limited (Clark and Root, 1999; Kumaraswamy and Zhang, 2001; Greenaway et al., 2004; Flinders, 2005; Aziz, 2007; UNECE, 2007; Garvin and Bosso, 2008). We agree with Rachwalski and Ross (2010) that more work is needed that looks at how governments might best organize their PPP programmes.

We argue that a shortcoming of the PPP literature to date is its normative approach to the subject, creating, in effect, a ‘one size fits all’ view of PPP programmes. For this reason, there remains a need for work that recognizes the way that PPP programmes are implemented differently in different regions. To enable this approach, we employ the concept of a ‘PPP-enabling field’ (see Jooste and Scott, 2011).

The emergence of the ‘PPP-enabling field’

There is a growing understanding among scholars that the move towards private participation in infrastructure does not simply substitute private sector capacity for public sector capacity; rather, it requires that new forms of public sector capacity be developed to overcome various challenges that infrastructure PPPs face (see, for instance, Dutz et al., 2006). In previous work (see Jooste and Scott, 2011), we have shown that this PPP-‘enabling capacity’ has not been answered by a reformation of public agents alone—rather a network of new ‘enabling organizations’ (public, private and non-profit) has emerged. These organizations, in varying ways, attempt to support the development and continued operation of PPPs for the benefit of public, private and civic actors. These ‘PPP-enabling organizations’ include: Sponsoring Departments, PPP Units, Transaction Advisors, Transaction Auditors, Public Regulators, Non-public regulators, Advocacy Associations and Local, Regional and Multinational Development Agencies.

We have proposed that the concept of an organizational field (DiMaggio, 1991; Scott and Meyer, 1991; Scott et al., 2000), found within institutional theory, can be usefully employed as a theoretical lens in this regard. DiMaggio and Powell (1983, p. 148) defined an organizational field as ‘those organizations that, in the aggregate constitute a recognized area of institutional life: key suppliers, resource and product consumers, regulatory agencies, and other organizations that produce similar services and products’. We have termed this network of PPP-enabling organizations the ‘PPP-enabling field’ (see Jooste and Scott, 2011). This enabling field differs from others that might be identified, for example, the organization field (dominated by private providers) that arises to construct, oversee and operate the private side of a specific PPP project.

We believe the concept of ‘field’ is more helpful than that of ‘programme’ in that it draws attention to both the organizational and institutional aspects of PPP implementation in a region and allows consideration of the broader political and societal environment affecting the conception and design of PPP programmes.

Empirical puzzle

In previous work (Jooste and Scott, forthcoming) we examined the characteristics of three leading PPP-enabling fields: BC, Canada; Victoria, Australia; and South Africa. We specifically chose these cases for their institutional similarity—as British Commonwealth, Anglo-Saxon parliamentary democracies, we anticipated finding quite similar field structures. Our findings, however, revealed striking differences in a number of aspects, including the structure and institutional location of the PPP unit, the roles that different enabling organizations took in the field and the cognitive frames (or institutional logics) that guided action in each field.

Explaining this surprising variation of enabling field structure in quite similar institutional settings is the motivation for the current paper. We believed that examining the historical trajectories of how these fields were constructed might shed light in this regard. Specifically, we propose to show that PPP field development does not follow a single set path towards a unified programme model. Rather, programmes are shaped by the institutional and political contexts in which they are constructed and bear the markings of the political actors that take a hand in their crafting.

Overview of this paper

This paper offers a comparative analysis of the development of PPP-enabling fields in BC, Victoria and South Africa. We start off by recognizing (and briefly exploring) the wider context of public sector reform, of which PPPs played a part. We emphasize the important role played by the UK's PFI as the first ‘PPP-enabling field’ and an exemplar for other countries, particularly those in the UK orbit of influence. Thereafter, we describe the evolution and structuration of the PPP-enabling fields in our three cases. In this way, we show the creation of our case fields as influenced by both transnational and by more localized events and actors. Our analysis is based on semi-structured interviews with leading actors in each of the regions and on archival materials, including histories. We draw heavily on theories from institutional change and structuration in this regard. It is to these theoretical foundations that we turn to next.

Theoretical underpinnings

Before turning to our study of PPP-enabling fields, we provide a brief overview of our underlying theoretical conception. We rely on a combination of institution and structuration theory.

Institutions and structuration theory

Social scholars have long been interested in ‘institutions’—i.e. the symbolic elements that create shared meanings and controls that provide order to social action. These elements include regulatory and legal frameworks, norms and value systems and cultural elements such as schemas and beliefs (Scott, 1995, 2008). The study of institutions is helpful for our current work on the formation of PPP fields. It expands the focus beyond the explicit, intendedly rational aspects of the PPP construction to incorporate elements of the political and social context in which these fields were constructed.

Institutions are socially constructed frameworks of symbols and resources that both enable and constrain social action (Scott, 1995, 2008). Although they are sources of stability, they themselves undergo change, and much attention has recently been devoted by theorists and researchers to the nature of institutional change. Thinking about change processes has been advanced by the theoretical work of Giddens (1979, 1984) who developed ‘structuration theory’: a conception of social structure and its relation to social actors. In this formulation, social structure comprises two elements: symbolic systems (institutional elements) and material systems, including both human and non-human resources. Symbolic structures give meaning to resources while resources are required to build—both to reinforce and change—symbolic structures. Social structures are both the context for and the product of the activities of social actors. If social structures are to persist, they must be enacted by social actors; if they are to change, actors are the agents of change. In this way, structuration theory helps to connect macro-structures and forces with individual actors and the actions they take.

Individuals vary in their ‘agency’—the capacities they have to introduce change into any system. Depending on their natural endowments, their access to resources, their relationships (social capital), some are in a better position to induce change than others. Of particular importance are the ‘entrepreneurs’—those able to assemble resources into new combinations of material resources (Schumpeter, 1961) and ‘institutional entrepreneurs’ (DiMaggio, 1988)—those actors capable of combining symbolic resources into new frameworks supporting and guiding action.

As noted, social structure provides the context for action. All actors are situated in specific contexts bounded by time and space and, as a result, social structures are ‘sticky’—resistant to change. Most social change is a ‘path-dependent’, incremental process (David, 2001), as later developments build upon and are shaped by earlier developments. Change can occur through developments internal to the sector or because of broader forces operating in the environment of the field. Being symbolic, institutional elements are readily transported from one social arena to another.

Mechanisms

Scholars have identified a number of specific ‘mechanisms’ associated with change. Mechanisms focus on the process of change—on how certain effects are produced. For example, Stinchcombe (1965) has identified the mechanism of imprinting—observing that populations of organizations frequently are founded in spurts, each cluster exhibiting similar structural features, reflecting the limited range of material and cultural elements available at that time. Some actors serve as carriers that diffuse institutional elements and practices from one place to another. Ideas—models, ideologies, policies—are readily transported by agents via the media and relational networks (Scott, 2003). While such mechanisms often result in the spread of similar institutions, creating ‘isomorphic’ or ‘convergent’ change across organizations or fields, diffusion may also involve other mechanisms that create ‘divergent’ change. When institutional elements diffuse, they may be variously theorized or translated by their carriers or recipients (see Strang and Meyer, 1993; Czarniawska and Joerges, 1996). To be carried, cultural models must be simplified and codified, and those who do this work are likely to transform the meaning of the models transmitted.

An additional mechanism of change is termed bricolage (Douglas, 1986; Campbell, 2004, Chapter 3) and refers to the creation of novel combinations of existing elements. Bricolage can also work in connection with translational processes to combine new externally derived elements received through diffusion with existing local beliefs and practices. Such practices ease the acceptance of new, foreign elements into varying environments and hence foster institutional change.

We illustrate the application of some of these ideas regarding institutional change and structuration as we consider the case of the introduction and spread of PPP-enabling fields in recent decades. We draw specific attention to the interplay between symbolic systems (or the institutional context) and material systems (including actors, networks, artefacts) and the mechanisms of change (specifically diffusion, translation and bricolage) that this interaction represents.

Method

Our research design is a multiple-case study (Eisenhardt, 1989) of three leading PPP fields: in BC (Canada), Victoria (Australia) and South Africa. Although these regions have long histories of public and private interactions, our focus here is on the recent history of PPP field development (in all three cases after 1990). Time and space requirements preclude our providing an in-depth historical account in this regard, so we have attempted to supplement our high-level review with selected explanatory footnotes while citing references that provide further detail.

Although our broad conceptual framework employs concepts and arguments based in institutional and structuration theory, we employ grounded-theory building approach because of the lack of previous work on PPP field development (Glaser and Strauss, 1967). The multiple-case approach is useful as it helps us to abstract results that are more generalizable (and therefore reliable) than those drawn from a single-case study (Yin, 2003; Taylor et al., 2009).

We selected our cases based on three factors: (1) all three cases are widely recognized as leading examples of PPP-enabling fields (Davies and Eustice, 2005; Cuttaree, 2007; PPIAF, 2007; Farrugia et al., 2008); (2) the cases present a large variation in the ideological foundations of the ruling political party and (3) as British Commonwealth, Anglo-Saxon parliamentary democracies, these cases provide at least some control for external variance in the underlying institutional settings, thereby simplifying the comparative analysis. In this way, these cases help us to theorize about the nature and trajectories of PPP field development.

For our analysis, we drew from three separate but complementary data sources. Our primary source is semi-structured interviews with leading actors in both public and private organizations in each of the case-study regions. We selected informants from a range of organizations intricately involved within the PPP field, including the PPP units, sponsoring departments and private advisors. Initial informant selection was based on recommendations made by leading representatives in the respective PPP units. This was supplemented by a ‘snowball’ approach where each informant suggested other salient field actors that we should interview. In this way, we triangulated the suggestions of the PPP unit representatives. We conducted a total of 42 interviews with 50 informants distributed evenly between the three regions and the enabling organization types. Confidentiality requirements preclude the publication of the details of informants, but we include summary information of the informant list in Table 1.

Table 1 Distribution of interviewees

Enabling organization	BC	Victoria	South Africa
PPP unit	6	4	1
Sponsoring department	3	6	4
Financial advisor	2	2	2
Legal advisor	1	4	1
Other consultant	1	1	1
Auditor general	1	1	1
Advocacy organization	1	1	1
Fairness advisor	1	1	0
Development agency	0	0	2
Federal PPP unit	1	0	0
Total	17	20	13

Interviews were recorded, transcribed and then coded in a systematic iterative manner with the use of the qualitative coding software QSR Nvivo. This software is specifically useful for drawing out trends from large amounts of qualitative data (Bazeley and Richards, 2000). We supplemented our interview data with other data sources, including documents and secondary data that were either publicly available or provided by our informants, to increase the validity of our findings (Eisenhardt, 1989). We complemented and triangulated our interview data with a review of existing literature on PPPs in the three regions. This included academic articles, reports, newspaper articles and online data sources. The bulk of the analysis was, however, based on interview data. Lastly, we also had our informants review early drafts of this article to ensure the accuracy of the findings.

Origins of the PPP-enabling field

This paper focuses primarily on divergent change. Still, we would be remiss if we did not recognize the overlaying process of convergent change that ushered in the use of PPPs for infrastructure delivery in the cases under question. Indeed, the fact that our case regions all implemented the use of PPPs since 1990 could easily be framed as narrative of isomorphism (the diffusion of structurally similar forms). Before we consider the details of field formation in each of our cases, we therefore first need to briefly consider both the broad reform trends which ushered in the consideration of PPP-type arrangements and the emergence of an enabling field ‘archetype’ that could be diffused more specifically.

Overarching reform trends

The involvement of private enterprise in the delivery of public services is not new. In fact, some have traced back this practice to previous centuries (see, for example, Garvin, 2007) and even ancient times (Ghobadian et al., 2004). The last three decades have, however, seen a striking resurgence of this approach, leading to widespread and deep reforms in the way that public services are delivered globally. Reforms driven by perceived public sector inefficiencies and the ascendancy of liberal economic ideology (Batley and Larbi, 2004) entailed a move towards ‘New Public Management’ (Hood, 1991; Osborne and Gaebler, 1992; Rhodes, 1996) and the adoption of neo-liberal economic policies (Henisz et al., 2005).

PPPs represent a late incarnation of these reform movements.1 Having aggressively pursued privatization of large parts of the public sector, many governments were looking for a new way of engaging with the private sector (Broadbent and Laughlin, 2003). This was specifically the case in the UK, where the government hoped to move from selling off existing state assets to an approach that entailed the development of new assets in ‘partnership’ with the private sector. In this way, the UK government hoped to overcome public financing constraints (Clark and Root, 1999; Parker, 2009). Although PPP-type projects were introduced in varying intensity elsewhere prior to 1992, the establishment of the PFI is widely regarded as the birth date for systematized programmes aimed at encouraging PPPs (Allen, 2001; Ghobadian et al., 2004). As such, it presents the birth of the concept of the PPP-enabling field.2

PFI: birth of a concept

The Conservative government of John Major launched the PFI in the autumn of 1992. By 1993, the precursor of a PPP unit emerged with the establishment of the Private Finance Panel, a group of high-ranking private and public officials whose role it was to encourage greater PFI deal flow (Allen, 2001). The next five years, however, saw only limited implementation of PFI projects (Broadbent and Laughlin, 2003). Clark and Root (1999) provided a detailed review of the difficulties that the PFI programme faced in getting off the ground during this time, including ideological opposition, misconceived notions on the ease of risk transfer and a resistance by local councils to the Major government's ‘central control of the PFI process’ (Clark and Root, 1999, p. 356).

When Tony Blair's Labour government came to power in 1997, many anticipated that it would be the death knell for PFI—indeed, in opposition, they had been severely critical of the programme (Flinders, 2005, p. 217). Instead, Labour set about giving the PFI a substantial facelift. This included rebranding it with the ‘warmer partnership ethos’ (Hodge and Greve, 2005, p. 310) by introducing the term ‘PPP’ (Wettenhall, 2005). In addition, the Labour government radically altered the PPP field, most notably by establishing the first PPP unit within Treasury, known as the Treasury Taskforce (Allen, 2001). This greatly decentralized the implementation of PPPs, with the PPP unit viewed as a ‘guardian of policy principles and promoter of best practice’ that worked in parallel with ‘highly skilled Departmental procurers pursuing their projects without central support’ (HM Treasury, 1997, p. 2, quoted in Clark and Root, 1999, p. 357).

After three years (in 2000), the taskforce was replaced by a permanent organization, Partnerships UK (PUK). PUK was itself formed as a partnership, with the organization formed as a Public Limited Company that was majority owned by private investors (51%, while government owns the remaining 49%). In addition, PUK was to be funded from project fees on a non-profit maximizing basis (Farrugia et al., 2008). The part of the Treasury Taskforce responsible for policy formation was taken over by the Office of Government Commerce in Treasury.

We note here that PPP projects in the UK have to date remained ‘off-balance sheet’. The future liabilities (of unitary payments for the length of the concession) do not show up on government books (Allen, 2001). This has continued to draw criticism in the UK (see, for instance, Flinders, 2005). Other notable aspects include a high level of contract standardization between PPP deals and the retention of core services (such as clinical services in hospitals, education services in schools and custodial services in prisons) by the public sector.

Identifying typical carriers

The PFI created a symbolic system that emerged as an archetype of PPP implementation, forming a seedbed of ideas that were carried through material systems to each of our three case regions. Far from being a one-time transfer, this diffusion continued as fields were formed and reconfigured over time. This diffusion relied on three main material systems or carriers to transport these institutional concepts across borders:

• Relational networks: PPP ideals were conveyed along relational networks between actors in each of our case regions and actors in the UK. The Anglo-Saxon roots of our three cases made these relationships particularly influential, so the rapid diffusion should not be surprising.

• Actors: A number of actor types emerged as carriers. First, local advocates for private participation in infrastructure embraced the developments abroad, and so supported their local adoption either informally or through more formal ‘advocacy organizations’ (Jooste and Scott, 2011). In addition, diffusion was facilitated by consultants employed by local governments in search of solutions to infrastructure problems. These consultants most prominently included UK expatriates who had participated in early PFI projects and were employed in temporary or more permanent capacities. Thirdly, multilateral and international development agencies propagated the use of PPPs and the formation of PPP-enabling fields as one component in a wider normative drive for the adoption of neo-liberal economic reforms. This carrier group notably included Bretton Woods organizations (the World Bank and the IMF) and other actors in the donor community (for example, USAID and GTZ).

• Artefacts: We recognize the way that artefacts (Scott, 2008, pp. 83–85) served as carriers in the diffusion process. We refer, for example, to various PPP documents such as project agreements, policy documents and implementation guidelines. These documents, embodying standardized practices in the UK PPP-enabling field, travelled to our case regions when local actors searched for baseline documents in the formative days of those fields.

Although these typical carriers can be identified in aggregate, diffusion was not a uniform process across our three cases. As we will show below, different combinations carriers were involved in each case. Moreover, processes of translation and bricolage ensured that the diffused symbolic systems were reconstructed very differently in each case. These historic accounts of field construction are elaborated below.

PPP-Enabling field formation in our three cases

We already alluded to the fact that the formation of PPP-enabling field entailed aspects of both convergent and divergent change. Although many of the PFI-archetype's structures and practices were diffused to each of our cases, they were subject to local translation and editing and were combined with existing institutions in a process of bricolage. In order to paint a more detailed historical picture of each field, we consider each case in turn. We draw specific attention to selected aspects of the institutional context (specifically, the political context at the time of formation) and the role of the carriers and mechanisms we identified earlier.

British Columbia

The use of PPPs as an infrastructure procurement method was not introduced in BC until Gordon Campbell's (fiscally conservative) BC Liberal Party came to power in 2001. An examination of the institutional context, specifically the political conditions in BC leading up to this election, is important for understanding the introduction of infrastructure PPPs at this point in time.

During the 1990s, tax payer debt increased rapidly under the (pro-labour) New Democratic Party (NDP) administration. This was due in part to the NDP's reluctance to raise taxes, a failure to recognize the extent of the accumulated debt burden (addressed later by the introduction of the Consolidated Capital Plan) and a penchant for state-centric development (Milke, 2001). The latter was epitomized by what came to be known as the ‘Fast Ferries scandal’ (Stewart, 2008)—an ultimately unsuccessful governmental project to design and construct a fleet of high-speed ferries, aimed at reviving the shipbuilding industry in BC. Beyond revealing the NDPs ideological foundations, Fast Ferries also highlighted the left's longstanding ‘love affair’ (Milke, 2001, p. 25) with organized labour, including the public sector unions engaged in infrastructure delivery. It is this bond that prevented the NDP from introducing PPPs for infrastructure development, even after it had investigated the UK's budding PFI programme. BC's deep fiscal problems and the ferries controversy, among others, culminated in the most one-sided election victory in the province's history: the Liberals won all but 2 of the 79 seats in the provincial legislature.

This political context alone was, however, not sufficient for the emergence of PPPs, but rather relied as much on the actions of Campbell as central field actor (or institutional entrepreneurs). Using the political opportunity presented by the strong voter mandate, he implemented extensive fiscal reforms (such as the sale of a number of state assets, various tax cuts and serious cutbacks in the size and budget of the civil service), including the introduction of PPPs as a mechanism for large infrastructure delivery. One of our informants (a leading financial advisor) explained it as follows:

And I think what Campbell was able to do particularly in the early days was to say ‘well you know, I'm basically the government.’ In his early [days] he had all but [two seats] so he was able to do whatever he wanted to do, and he wanted to change government, and he did it at all kind of stages.

The introduction of PPPs in BC followed a very purposeful adoption of the UK model, and as such presents the clearest instance of diffusion in our three cases. This diffusion was enabled by a combination of relational networks, local advocates and artefacts. Representatives from BC's Provincial Treasury had undertaken a reconnaissance trip to the UK to study the PFI as early as 1994, and they conducted a more comprehensive investigation soon after Liberal Party leader Gordon Campbell came to power. Local advocacy through the Canadian Council for PPPs lent additional support at this time, with this non-profit organization even flying in specialists from the UK to support ministers attempting to launch pilot projects. The following quote by a representative of the BC PPP unit, unsurprisingly called Partnerships BC (or PBC), mimicking the UK group called PUK, highlights the purposeful diffusion process that was followed in establishing the PPP-enabling field:

We looked to them [Partnerships UK] a lot. I think we started with Partnerships UK's Standard Concession Agreement and have kind of adapted it to ours. We have gone over and done tours of projects and met with people to learn how they do things and that's with Partnerships UK.

The later emergence of PPPs (2002) in BC meant that documentation had evolved very far by the time that the documents were diffused to BC. In addition, British consultants readily travelled across from the UK to work on pilot projects in BC (often in association with local firms). The variety of carriers and their reception led to the BC field having the closest resemblance to the UK model among our three cases.

Even though this introduction of PPPs so closely adopted the UK model, we can observe some forms of translation that took place. This is specifically reflected in two main differences between the BC field and the UK archetype. First, the field was infused with a centralized top-down approach, as evident from the way that PPPs were implemented. Two of the three pilot projects undertaken faced significant sponsoring department resistance and only went ahead when the provincial government refused to consider non-PPP delivery (for details, see Cohn, 2008, pp. 78, 79). The development of the Capital Asset Management Framework (effectively the guiding PPP legislation) gave central priority to PPP delivery: public officials were required to show why a project should not be done as a PPP. As Cohn notes, this policy ‘turn[ed] the tables on the procurement process, making it necessary to justify not using mechanisms such as [PPPs] rather than forcing public servants to defend their use’ (2008, p. 92). In addition, some of our informants indicated that Campbell directly incentivized newly appointed ministers within sponsoring departments to pursue PPPs in the early days of the PPP programme. We have argued elsewhere (see Jooste and Scott, forthcoming) that this centralized top-down approach led to the emergence of a ‘logic of enablement’. This is especially reflected in the role that PBC took on projects, driving project delivery rather than only overseeing it (as was the case in Victoria and South Africa).

Second, the translation process reflected the Campbell administration's opposition to organized labour. From the outset PPP projects in BC followed what can be described as a ‘puritan approach’ to the allocation of services between public and private parties. Many activities that were retained in other jurisdictions (such as cleaning and portering in hospitals3) were outsourced with apparent disregard to the strong public sector unions that dominated some of these industries. This has contributed to criticism levelled against PPPs by many of these public unions have remained critical of PPPs, as has the pro-labour NDP. As a result, PPPs have continued to be a politically polarizing issue, as is evident by the airplay it received during the 2009 BC elections (which Campbell's Liberal party won). The BC government has softened this puritan approach somewhat in more recent health projects, where some of these services have now been excluded from PPP delivery.

Victoria

We have already alluded to the fact that Victoria has a much longer history of private participation than BC. This history of the PPP-enabling field can be divided into two periods (English, 2006; see also Maguire and Malinovitch, 2004): (1) early field development under the Kennett government and (2) a substantial revision under the Bracks government.

However, reformation of the field in the second reveals many more traces of diffusion, as we show below.

Pre-2000: early formation under Kennett

Our investigation of the Victorian field construction again starts with a limited consideration of the institutional context into which it took place.4 From the late 1980s until 1992, the ruling Australian Labor Party used PPP-type arrangements (not full PPPs in our definition, due to the very low-risk transfer taking place) as a way to achieve off-balance sheet financing to circumvent the limits set by the Australian loan council. This off-balance sheet financing contributed significantly to a burgeoning of public debt levels, up to around $30 billion when Jeff Kennett's (fiscally conservative) Liberal Party took over in 1992. As was the case with Campbell in BC, the fiscal failures of the Labor government not only paved the way for a landslide victory for the Liberals, but presented Kennett with his own political opportunity for reform.

Kennett represents a similar central field actor in this early field construction process. Between 1992 and 1999, the Kennett government implemented a fierce and radical programme of budget-cuts and fiscal reforms, which he himself dubbed the ‘Kennett Revolution’ (Parkinson, 2000). This included the privatization of a number of state services (most notably the gas and electric utilities) and the construction of various large-scale infrastructure projects under early PPP arrangements. These first pilot PPP projects included the CityLink highway project, various water and wastewater treatment plants, the Latrobe and Mildura Hospitals and a number of prison PPPs.

The relatively early emergence of PPP type projects under Kennett (almost in parallel with developments in the UK) means that a diffusion account does not so readily explain the emergence of this early PPP-enabling field. Not surprisingly, these early projects show substantial diversion from the UK model in at least three important respects (Maguire and Malinovitch, 2004). First, in response to the debt burden that resulted from prior projects, the Victorian Auditor General revised the tax treatment of PPPs, recognizing debt commitments ‘on balance sheet’. Second, projects entailed a full transfer of services from the public to the private sectors (even more so than the UK practice at that time). Some of our infomants suggested that this was done in an attempt by Kennett to break the power of public sector labour unions. A third notable aspect of PPPs under Kennett was that asset ownership for social infrastructure was transferred to the private sector (English, 2005) for the length of the PPP contract. This practice also differed from the UK's PFI projects from this time.

We lastly note that these projects were pursued under a very limited PPP policy (known as the Infrastructure Investment Policy of Victoria) and framework, although this resembled the approach taken in the PFI. As one informant (a treasury employee under Kennett), put it:

prior to 2000 it would be fair to say our guidelines weren't particularly strong and we just went out and did projects.

Post-2000: reform of the field under Bracks

By the end of Kennett's second term, the reform programme had begun to draw widespread criticism. The primary objection was the severity of the reforms (specifically the erosion of public sector jobs). Opponents also pointed to the limited rigour applied in project development (Maguire and Malinovitch, 2004; Hodge and Greve, 2005) and the lack of transparency and perceived probity issues surrounding government procurement at this time (exemplified by the Crown Casino project).

These aspects of the institutional context paved the way in 1999 for the unexpected election victory by Steve Bracks' Labor Party. Bracks had replaced long-time opposition leader John Brumby as Labor party candidate shortly before the election, as he was seen to provide ‘a softer face’ to Kennett's reformist approach. Some feared that a Labor victory at this point would signal the end of PPPs in Victoria, but, as had been the case in the UK, Bracks' Labor government set about amending the model, and then expanding the PPP field. The role of Brumby as central field actor should not be overlooked here. In fact, in many respects he emerged as the main proponent for the continued use of PPPs. One of our informants, who was intimately involved in amending the PPP model, explains:

I spent two years in John Brumby's private office advising him at that time as finance minister but he soon became the treasurer and much later premier. Back then, as now, he was the major political supporter and champion of PPPs in Victoria. [The] PPP policy was John Brumby's policy and he gets the credit for driving it forward and realizing the many benefits that it has brought to infrastructure development in Victoria.

The amendments that Brumby implemented during Bracks' first term were in direct response to the shortcomings of PPPs under Kennett. These amendments revealed a number of traces of diffusion, with expatriate consultants from the UK serving as specific carriers. In fact, the informant quote above was such a consultant. Having built up considerable experience on PFI projects in the UK, this informant was employed by the Bracks government to help reformulate the PPP policy. In an interview, he described how they used the ideas of a central player in the PFI to structure their field:

That's why I say we learned from the UK because … he [the central player] was taking a look back and saying, that's what you need: political will; focus on value for money; rules and guidelines; training; and of course projects. And [even] they [the UK] didn't have all of them. And so, we just looked at that as a model to adopt, because he obviously knew what he was talking about and it makes intuitive sense.

In addition, artefacts have served as carriers, as field actors have continued to rely on project documentation from the UK. The early local experimentation with PPPs in Victoria, however, meant that the diffused practices from the UK were reconfigured and combined with local models, exemplifying a process of bricolage.

Four changes to the PPP-enabling field implemented by Brumby are worth highlighting. A first step was to develop a clear PPP policy and to make it freely available. The Partnerships Victoria (PV) policy as it was coined, adopted in 2001, was widely touted as the foremost international PPP framework at the time. It included the adoption of project development processes that had become standard in the UK (including the development of a Public Sector Comparator and a Value for Money assessment) while adding new innovative practices (such as the formal public interest test). In this way, the Bracks government attempted to address the dual criticisms of a lack of development rigour and transparency. An informant who was involved at the time explains:

The guidelines were written and … they were much more comprehensive than they ever were and much more detailed and all the rest. But I think in part that was also allowed the Labor government to demonstrate how its approach was to be more transparent.

In addition, in order to get buy-in from the wider Labor party, Brumby (and Bracks) set in place a more balanced contracting-out of services in PPPs, an approach that was non-threatening to public labour unions. Not only were core clinical and custodial services retained in hospitals and prisons, respectively, but on many projects non-core services such as cleaning were also retained by government. This change widened the definition of ‘core services’ beyond that set out in the British model.

A related change was an attempt to distance the PV field from a term that had grown increasingly contentious: privatization. This was done by taking back the ownership of assets under PPP deals. Together with the retention of core service, this change was aimed at presenting a ‘softer face’ of PPPs. To this day, the Victorian government (and proponents of PPPs in Australia in general) is at pains to highlight the distinction between PPPs and privatization.

A final noteworthy change that the Bracks government implemented was the reformation of the PPP unit within Treasury. Although a similar unit had existed under Kennett, the new unit took a much less hands-on approach to project implementation, as it was mainly tasked with overseeing the implementation of the PV policy. An informant, who was employed by Treasury at the time, explains:

When I was in treasury [under Kennett] I was leading the project and treasury ran up the project and then gave it back to the department. This government [under Bracks] said ‘Well you can't sort of force things down people's throat,’ so they changed that.

Clearly, this change was implemented both to increase buy-in by Sponsoring Departments and to convey the ‘softer face’ of the PPP field, as noted earlier. This change mirrored to a large extent the decentralization that had taken place under the UK's Labor government. The PPP unit that was formed took on a similar role to that of PUK, but its institutional structure was closer to that of the Treasury Taskforce (PUK's predecessor).

The revised approach by the PPP unit also led to other instances of bricolage, as existing organizations adapted to fill new roles in the field. A prime example is Major Projects Victoria (MPV), a specialized project delivery unit within the Department of Innovation, Industry and Regional Development. MPV had been established in the Kennett years, when government set about delivering a number of large high-profile infrastructure projects (not under the PPP regime). With the move to a more decentralized PPP delivery model under Bracks, some departments were not able to build up the necessary project delivery capacity needed to develop and deliver projects immediately. MPV's existing expertise in project delivery made it the natural choice to assist low capacity departments in this regard.

Both the Victoria and BC cases entail the centrally coordinated and purposeful construction of a PPP field (albeit through more than one iteration in the Victorian case). Not all PPP fields emerge in this way. Some fields, it would seem, follow a much more evolutionary path, with various streams progressing in parallel at times, before merging later. Such was the case in the development of the South African PPP field.

South Africa

Our examination of the construction of the PPP-enabling field in South Africa again starts with an evaluation of the institutional context into which it emerged. The first democratic government in South Africa came to power in 1994 with the historic triumph by Nelson Mandela's African National Congress (ANC). It did so in a ‘Tripartite Alliance’ with two other major political organizations, the South African Communist Party and Congress of South African Trade Unions. The new government faced the challenge of addressing significant infrastructure requirements (to redress imbalances created by Apartheid) with an alarming budget deficit inherited from the outgoing National Party. In addition, discrimination against blacks under Apartheid meant that most government departments were struggling to resource managerial positions, causing ‘a general lack of capacity to manage facilities and provide required levels of service within the public sector’ (Merrifield et al., 2002).

A striking feature of the South African case is the absence of a central field actor or institutional entrepreneur who enabled the introduction of PPPs. Rather, a number of more peripheral actors had a hand in their emergence. Most prominently, a number of international development agencies served as advocates, applying normative pressure on South African government for the use of PPPs (as part of their neo-liberal policies) as a solution to these intractable infrastructure problems (Niksic, 2004). The Tripartite Alliance's socialist roots, however, prevented a central PPP initiative from developing. Rather, governmental departments started to investigate PPPs in a somewhat decentralized manner. This was the result of more focused lower-level influence that came through technical assistance projects undertaken by the donor community and international consultants.

Three main organizations began to implement PPPs quite independently at this stage: the South African National Roads Agency (SANRAL), the Department of Correctional Services and the Municipal Infrastructure Investment Unit (MIIU). We discuss each of these PPP-enabling field development streams in turn.

PPP stream 1: SANRAL

The first governmental department to attempt PPP-type projects was the Department of Transportation, which started investigating their use for national roads in the early 1990's (at that stage still under the National Party government). This was done predominantly as an attempt to move infrastructure spending off the department's balance sheet. This culminated in 1997 in the first PPP-type project in South Africa, the N1/N2 project. Because this was prior to the development of any PPP legislation or methodology by National Treasury, legislation specific to national roads was passed in 1998 in the form of the SANRAL act. This act provided for the establishment of a public corporation (similar to PBC) responsible for managing the construction, operation and maintenance of national roads. The act further allowed this corporation to enter into PPP-type contracts with private providers and levy tolls on National roadways.

Further projects followed in the form of the N3 toll road, N4 Maputo Development Corridor and the N4 Platinum Highway. By the time that the PPP Unit was established and PPP guidelines published (see below), SANRAL had been so successful at building up PPP delivery capacity, that National Treasury effectively excluded them from the PPP Unit's oversight. To this day, national road PPPs remain separate from the rest of the South African PPP field.

PPP stream 2: Department of Correctional Services

Up to the late 1990s, social infrastructure assets (such as hospitals, schools, prisons and offices) were delivered by the Department of Public Works (DPW) on behalf of various sponsoring departments. Towards the late 1990s, DPW was acutely aware of the challenges that sponsoring departments were facing in terms of growing infrastructure requirements under declining budget allocations. This was specifically salient for the Department of Correctional Services (DCS), where the provision of accommodation for a growing prisoner population required urgent action. As a solution, DPW contracted with British consultants to develop a framework for implementing PPPs for social accommodation in the South African context. These consultants, along with the artefacts they brought along (including PFI policy documents and standard contracts), served as carriers in a very direct diffusion of the PFI archetype. The resulting Asset Procurement and Operations Partnership System (APOPS), a step-by-step PPP guideline document, very closely resembled the PFI guideline documents in use in the UK at that time (Merrifield et al., 2002).

DPW's implementation of APOPS for DCS culminated in the identification of preferred bidders on six proposed PPP prison projects. An informant in treasury takes the story further:

It was at a very late stage in the whole process that they came to treasury and said, ‘It's going to cost us this much.’ This means, on an annual basis they had to pay what we call a unitary payment. And that unitary payment would have consumed on these 6 facilities a significant part of their existing budget. They would actually have to close down other facilities. So Maria Ramos who was the DG at the time was very clear that the system could not work this way. You can't have departments going out, procuring PPP's, pretending that they have no fiscal implications because there is no capital investment by the state, because you incur this long term obligation that impacts their budget. It became very clear that there was a very strong role for Treasury to play in regulating the financial aspects of these [projects].5

This chain of events became the main driver for the subsequent development of a central PPP-enabling field in South Africa. This entailed instances of both additional diffusion from the UK archetype and significant translation of these diffused elements. First, the PPP unit that was subsequently established within National Treasury in 2000 entailed a close diffusion of the UK model. An informant from the South African PPP unit explains:

And I think that we were lucky at the time that the UK was quite far advanced in terms of its PPP model and the Treasury task team had been through a learning curve and had set up systems that were quite replicable in South Africa. Similar judicial system, similar budgeting system. There was [sic] a lot of common elements there. … We sat down and we actually within eight months of sort of kick off, we had issued practice notes that weren't the most comprehensive and they were largely borrowed from what we saw then as being best practice … from the UK.

This diffusion relied on a number of carriers. First, two international development agencies (the German GTZ and the United States Agency for International Development or USAID) played important roles through direct technical assistance. In addition, strong existing relational networks with Britain meant that Treasury officials relied heavily on UK experience. Lastly, diffusion was also conveyed through the adoption of various artefacts from the UK.

But the controversy surrounding APOPS also led to significant translation. Specifically, the PPP unit was established in Treasury with a strong mandate to control the discretion of sponsoring departments in future. For this reason, the South African PPP field has continued to be imprinted by what we have elsewhere termed ‘a logic of control’ (see Jooste and Scott, forthcoming, to be contrasted with a ‘logic of empowerment’ that prevailed in PBC).

PPP stream 3: MIIU

The final independent PPP stream to emerge in South Africa was at the municipal level, where the challenges posed by building infrastructure under severe fiscal and capacity constraints were especially severe (Bahl and Smoke, 2003; MIIU, 2006). In this context, the Department of Provincial and Local Government (DPLG) started searching for solutions. The role of international development agencies (specifically the World Bank and USAID) in proposing the use of PPPs is again important here. Their support, along with that of a local public infrastructure investment bank (the Development Bank of South Africa), culminated in 1997 in the formation of the MIIU (MIIU, 2006; Smith, 2008). The MIIU was a government-owned non-profit organization ‘tasked with providing technical assistance and grant funding to municipalities investigating innovative [PPPs]’ (Hlahla, 1999, p. 565). Formed more than three years before the central PPP unit in Treasury, the MIIU acted as an early PPP unit in the municipal sphere.

The unit was partly funded by USAID and staffed by a combination of World Bank-appointed international consultants and local specialists. What should be noted is that the MIIU had no formal tie to National Treasury (Hlahla, 1999, p. 565), which at that stage was only starting to consider its role in PPPs (as noted above). It was hoped that this independence would help the MIIU ‘make deals happen, even in the face of taking some risks’ (Hlahla, 1999, p. 5). Clearly, the MIIU operated much more out of a ‘logic of empowerment’ than a ‘logic of control’ (see Jooste and Scott, forthcoming).

Over the next eight years, the MIIU was highly successful in developing projects, initiating over 100 projects on behalf of local authorities (Magugumela, 2005). The majority of these were quite small and do not fit our current definition of PPPs, but the list did include a number of large-scale PPP arrangements. These projects were implemented under the Municipal Systems Act (MSA) of 2000, legislation authored by DPLG, which allowed for PPPs in the municipal sphere. From the start, the unit was envisioned to only operate for a finite term, and by 2006, the organization was dissolved and its activities ‘taken over’ by the PPP unit in Treasury.

It is interesting to note that even before this formal dissolution, National Treasury attempted to obtain greater control of municipal PPPs through its implementation of the Municipal Financial Management Act (MFMA) in 2004. This added an additional layer of complexity to the contracting of PPPs over that of the existing MSA. In this way, it superimposed the requirements of Treasury (as reflected in the MFMA) onto those of DPLG (reflected in the MSA), again highlighting the salience of the ‘logic of control’ that has so imprinted the South African PPP field.

The PPP unit has more recently undertaken significant work to streamline the process of implementing municipal PPPs (specifically clearing up the confusion surrounding the duplicate requirements of the MSA and MFMA). This has, to some extent, led to a resurgence of interest in municipal PPPs.

Affirmative action: an instance of bricolage

The three streams illustrate the fragmented nature of PPP field construction in the South African case. In closing, we wish to highlight the way that the government's Affirmative Action policy has been combined with the PPP field structure in a process of bricolage to further shape the field structure. Broadly referred to as ‘Black Economic Empowerment’ or BEE, this affirmative action policy has been central to the ANC's ideology and strategy since it came to power in 1994 (ANC, 1994; Tangri and Southall, 2008). BEE entails an attempt to redress the inequalities created by Apartheid by affording persons previously disadvantaged (effectively non-white individuals) under the Apartheid policy the new economic opportunities. In this way, the ANC hoped to ‘de-racialise business ownership completely through focused policies’ (ANC, 1994). The ANC's parallel imperative of fostering rapid economic growth among the predominantly white owned business community has, however, limited the extent to which they have been able to enforce BEE requirements (see Tangri and Southall, 2008).

While the ANC government has been cautious to implement stringent BEE requirements on the general business sector, one area where they have seen considerable BEE opportunities has been PPPs. The government has recognized a number of reasons why PPPs present unique BEE opportunities (see National Treasury, 2004). These include: (1) the long-term nature of the arrangements presents unique opportunities for skills transfer to affirmative enterprises; (2) the fixed income stream typical of social accommodation PPP projects helps affirmative enterprises grow over time by decreasing business risks and (3) the high-profile nature of these projects make them valuable vehicles for showcasing the government's BEE achievements (National Treasury, 2004, p. 7).

As a result, the South African PPP field has been greatly shaped by the concept of BEE. Not only do BEE goals continue to form a central part of PPP bids requirements, but the achievement of BEE objectives is now one of the main selling points of undertaking PPPs in South Africa. This characteristic appears to be unique to the South African PPP field (certainly in our collection of cases, but possibly even internationally).

Conclusion

The aim of this paper has been to address an identified dearth of work on PPP programmes. Specifically, it has challenged a generally normative, ‘one-size fits all’ view of PPP programmes. To this end, we investigate the programme formation histories in three leading PPP fields, chosen because they have very comparable institutional settings. In doing so, we hoped to show that, even within similar contexts, programmes are implemented quite differently.

Our discussion leans heavily on our previously defined concept of the PPP-enabling field. This concept is useful in drawing attention to both the organizational and institutional aspects of PPP implementation in a region and allows consideration of the broader political and societal environment affecting the conception and design of PPP programmes. In addition, our explication of the development histories in these fields relies on a conceptual framework drawn from institutional and structuration theory. Specifically, we focus on the interaction of the institutional (particularly political) context, the actors and artefacts that were instrumental in shaping the field and the mechanisms by which this change happened.

We find generally strong support for our main thesis that PPP-enabling field develops very differently in regions sharing a number of common features. In addition, we illustrate the role that the particular institutional (and political) context and field actors play in bringing about the formation of the field and change in it over time. Specifically, we show the importance of a central field actor or ‘institutional entrepreneur’ in determining the development trajectory that the field follows. Both the Victoria and BC cases entailed a central institutional entrepreneur that took a leading role in coordinating and purposefully constructing the PPP-enabling field (in the Victorian case, this involved a substantial revision of the field as well). The South African PPP field lacked this type of central institutional entrepreneur, with the main early PPP proponents being non-local consultants and development agencies. This resulted in construction following a much more evolutionary path, with various streams progressing in parallel at times, before merging later. We also observe a number of change mechanisms, including instances of diffusion, translation and bricolage. These mechanisms lead to both convergent and divergent change in different aspects of our three cases.

There are two main contributions from this study. First, we draw attention to the need for a context-specific approach to explain and predict PPP field development. We believe this is especially important for governmental agencies proposing the development of a PPP programme. Second, we develop arguments relating to institutional systems and processes that can guide future studies of these and similar fields.

Nevertheless, this study represents only an initial step in investigating the development and evolution of PPP-enabling fields. Future work could look beyond our narrow institutional variety to include cases that did not entail diffusion from the British model or those that presented further diffusion of the adapted models in our three cases. In addition, it would be interesting to re-evaluate the structure of these fields at a later time, to see whether they attain a more stable form or continue to change with the winds of political change.

Notes

The exact ‘birth date’ of PPPs in our definition is somewhat unclear. Some have argued that the genesis lay in transportation projects that first emerged during the 1960s and 1970s in the form of tolled motorways in Spain, France, Japan and the USA (Ghobadian et al., 2004). We believe that projects that emerged in the UK and Australia during the mid-1980s bear closer resemblance (see also Broadbent and Laughlin, 2003; Parker, 2009).

It should be noted that similar PPP-type arrangements also emerged in Australia at this time. One-off projects in New South Wales were undertaken in the early 1980s. As we discuss below, the Kennett government in Victoria also pursued a dual privatization and PPP reform program through the 1990s. An even stronger case can be made for the actions of the Hong Kong government between 1987 and 1997 as it prepared for transfer to China. The ‘systematized’ use of build–operate–transfer type arrangements (Kumaraswamy and Zhang, 2001) can be viewed as one of the earliest incarnations of a PPP programme. The UK's PFI, however, represented a much more comprehensive and integrated roll-out of the PPP approach. As such, we recognize it as the first manifestation of a PPP-enabling field.

It is important to note that the legislation that enabled health authorities to contract out non-clinical services was introduced prior to the introduction of PPP agreements in health care. This legislation was introduced in January 2002, whereas the first hospital PPP agreement was only concluded in December 2004.

We recognize some similarities between the Kennett PPP implementation and that of Campbell (noted earlier), albeit nearly a decade apart. The Kennett reforms were, however, undoubtedly more severe, as they relied much more heavily on a full ‘privatization’ approach (highlighted by the transfer of asset ownership and full service delivery), while being pursued under a much less developed policy framework.

One of our other informants suggested that the reason for the over commitment on the part of DPW is to be found in their historical focus on project development, i.e. they are only concerned with the project development cycle and do not consider the actual operation and maintenance of assets. DPW therefore failed to appreciate the ongoing budgetary commitment these six projects represented.