Appendix 2: Auckland's Grafton Gully road construction
Procurement route
The new Grafton Gully road was constructed using a project alliance set up under new Transit New Zealand (Transit) policy. This procurement approach did not involve private financing or a transfer of ownership.
Grafton Gully connects the Auckland North-Western and Southern Motorways with the Ports of Auckland and lower Auckland central business district. The project was designed to reduce congestion in the area by separating motorway traffic from local traffic.
The 25-month project, completed in February 2004, cost $67 million.
Grafton Gully pioneered the project alliancing approach for road construction projects in New Zealand. The model adopted by Transit involves integrating the owner, designers, and constructors as a single team, sharing all risks and rewards and responsibility for solving problems.
The alliance team comprised Transit as the client (or “owner participant”), Beca Carter Hollings and Ferner Limited (a design consultant), and 2 principal construction companies (The Fletcher Construction Company Limited and Higgins Contractors Limited). A project alliance board was set up with representatives of each participant (including 2 from Transit, The Fletcher Construction Company Limited, and Beca Carter Hollings and Ferner Limited, and one from Higgins Contractors Limited), with a requirement that all decisions must be unanimous.
The model55 adopted by Transit was devised by an Australian consultancy organisation, Project Control International Pty Limited (PCI), with slight modifications to suit New Zealand’s conditions. Two main stages were involved in establishing the alliance.
Selection of alliance participants
In conventional procurements, price is generally the principal determinant of decisions about awarding the contract, although selection is often on the basis of both price and non-price attributes. In this case, short-listing of private sector proponents was undertaken based on management and technical skills of the organisations and the individuals proposed for the project, as well as the experience and track record of the organisations on similar work. Consideration was also given to how team members included in the project teams – proposed as part of each tender bid – demonstrated that they could work together effectively, and how management and technical skills they possessed would facilitate an alliance-type contract.
Two-day selection workshops were held, which were designed to enable the client and short-listed proponents to get to know each other, and to assess how they responded to quasi-hypothetical issues related to the project. The objective at the end of this process was to select the team with the most potential to deliver the project outcomes desired by Transit.
The project team selected at the end of this stage, which included representatives from both the client and successful proponents (known as the “non-owner participants”), became known as the “Freeflow Alliance”.
Determining price
Once the Freeflow Alliance had been set up, a process was then followed to establish a target outturn cost for the project. This process included 3 defined activities:
- Limb 1: An assessment of direct project costs, including project-specific overheads.
- Limb 2: A determination of the usual margin or direct project costs that each private sector participant would expect when undertaking normal business activities. An expert auditor was contracted to review the books of each private sector participant for the last 5 years, and to identify the normal margin for off-site overheads and profit that each earned on their normal business activities.
- Limb 3: Establishment of a method for the equitable sharing of “pain” or “gain”, depending on how project outcomes would compare with the pre-agreed targets which the parties would jointly commit to achieve.
Initially, the audit process determined Limb 2, and then the Limb 3 mechanism was prepared and agreed. The participants then entered into an interim Project Alliance Agreement (iPAA) to undertake sufficient design and construction planning to make the Limb 1 assessment.
This was then combined with Limb 2 described above to become the target outturn cost, including costs associated with all risks that were carried by the alliance – for example, fluctuations in labour and materials, and days lost because of adverse weather conditions.
Transit appointed an independent industry expert who used specialist staff to verify to its Board that the target outturn cost offered value for money. Any differences between the target outturn cost identified by the private sector participants and independent expert were to be resolved between them without Transit’s involvement.
Once the target outturn cost had been agreed, all participants signed the final Project Alliance Agreement (PAA), including provisions relating to governance by the Project Alliance Board, day-to-day management by the project team, sharing of risks, payments specified in accordance with Limb 1 to 3 assessments, and processes for resolving issues.
PCI, which has extensive experience in alliancing, facilitated the establishment of the core commercial arrangements that underpinned the alliance arrangement.
Reasons for procurement decision
This was a complex project, and it was imperative that a high-performance team was recruited to undertake it. This might have been difficult through a conventional competitive procurement process.
A traditional measure and value contract would not provide incentives to contractors and designers to “think smarter”, since they would not gain from seeking innovative solutions to cut costs. Under an alliancing approach, all participants would benefit from net profit gain.
The proposal for a non-adversarial approach was attractive to Transit, which saw advantages from the client and contractors not wasting time “covering their backs”.
Transit was aware that this approach had worked well in Australia for roading projects.
Some important lessons so far
Project team building
It was a major challenge bringing together 150 people from 4 very strong organisations to form a team that was going to deliver an outstanding project. This required a change in the behaviours and attitudes of people who worked in a very tough construction industry with a traditionally adversarial approach to contract management.
An Australian consultant specialising in alliancing arrangements was commissioned at an early stage to help establish the founding principles of the alliance, and a structured team-building programme was set up with 2 phases:
- Initial team building, coaching, and training sessions, and appointing a long-term coach.
- After the team had been together for 8 to 10 months, a review of progress and performance in setting up a fully integrated, high-performance team. A charter was written to capture the values of the team in everyday language, and initiatives were put in place to establish the culture and behaviour across the entire site.
Different mechanisms were also used to establish a team identity. For example, a single project office was established for project team staff, separate from the premises of their employers, with areas dedicated to the team relaxing away from their desks, and social occasions were arranged for the team outside work. Team-building activities were held to help foster the team identity and alignment on project objectives and “best for project” decisions.
The team was located on-site a month after the signing of the PAA, which focused the team on important outcomes and delivering the project.
An intelligent client
Transit considers that, in order for an alliancing approach to be effective, it is important for the public entity to possess expertise in this type of approach. Transit dedicated one expert member of staff full-time to the project, with 2 other staff regularly involved.
The alliance agreement required Project Alliance Board members representing the private sector participants to be empowered to make decisions on the part of their organisations. It was important that Transit staff could also make decisions without having to go back to Transit’s Board for ratification, particularly since responsiveness on the part of Transit provided confidence to private sector participants.
Time taken to establish the alliance
A great deal of work is required to establish an alliance – including agreeing the target outturn cost – as well as a need for ongoing team building. For the Grafton Gully project, setting up the alliance and agreeing to the target outturn cost took 4 months, although Transit has found that subsequent alliancing projects have taken longer to reach this stage.
Transit’s view is that, because of the “upfront effort” required, an alliancing approach would not be suitable for small projects, and that the Grafton Gully project was arguably too small. However, project size needs to be balanced against projects that might benefit from an alliancing approach, such as where a lot of stakeholders are involved, environmental issues are paramount, and superior outcomes are required.
External entities
The alliance identified a key area of risk as being anything influenced by a party or person external to the on-site alliance team. These included, for example, territorial local authorities, affected businesses and the community, and “home” organisations, including Transit’s corporate and regional offices. This risk was addressed through maintaining strong relationships with each of these external parties on issues that might affect the performance of the alliance.
Public relations
The close working relationship between the “stakeholder management team” and construction teams meant that construction activities that were likely to have a high public impact were identified and planned for well in advance. This had a positive effect on the quality and effectiveness of communication, leading to a high level of trust with the Auckland Regional Council and the emergency services.
Target outturn cost
There is a long period of uncertainty before the target outturn cost is known, and a risk that it will not be possible to agree the target outturn cost when considerable investment has already been made by all participants to reach this stage. For example, private sector participants and independent experts may have differences of opinion, which might lead to termination of the agreement.
In establishing a target outturn cost, Transit has found that assessments of private sector participants and independent experts relating to direct costs are usually closely aligned. However, costing of “preliminaries and generals” – for example, general project infrastructure and the support systems needed to deliver projects (such as project management staff, accommodation, IT, and vehicles) – and costing of risk factors can be more contentious.
There is also a risk that project costs will exceed the target outturn cost. In this circumstance the additional costs are borne by all participants in pre-agreed proportions (usually 50:50). The risk to the private sector participants will be having to work for reduced or entire loss of their profit and off-site overheads. The public sector participant will share the additional costs up to the point that the Limb 3 margin of the private sector participants is expended. After this, the public sector participant will have to bear all the additional costs. Also, there is always the risk that the reaction of private sector participants in this situation will be to revert to an adversarial approach.
Treatment of variations to the contract
In principle, there should be very limited possibility of variations to the contract in an alliancing arrangement that enable private sector participants to claim additional payments. Any variations should be limited to changes in the scope of the work requested by the client. Transit decided to organise a Variation Benchmarking Workshop to ensure that all participants in the alliance had a clear understanding of what would and would not constitute a variation. A number of hypothetical scenarios were discussed (such as changes to government legislation affecting working hours, and suppliers becoming insolvent) with an assessment of whether each scenario might result in a variation. The Project Alliance Board was required to make a unanimous decision on what would constitute variations, which were incorporated into the PAA as examples that could be relied on later if a situation arose.
Performance against programme
The project was completed 6 weeks ahead of the agreement completion date. This was achieved without compromise to the “non-production” objectives relating to environmental impact, traffic management, road safety, health and safety, stakeholder involvement, and quality and aesthetics, and despite a number of significant risks and obstacles that had to be overcome.
The alliance approach generated cost savings (mainly in relation to physical works) that were then fed back into the project as enhancements to urban design. For example, decorative panels, motifs on barriers, outdoor sculptures, and planting were incorporated into both the structures and site landscaping.
55: Introduction to Project Alliancing (April 2003), Jim Ross, Project Control International Pty Limited.