1. Executive Summary

In August 20201, the New Zealand Government asked the New Zealand Infrastructure Commission, Te Waihanga to oversee an urgent and wide-ranging Interim Review (IR) into the Transmission Gully Project (TGP). The IR's focus was on how the TGP agreement was awarded for the price agreed, whether this was a realistic price, and whether the risks identified were appropriate, properly considered and taken into account.

This Report outlines the approach undertaken, project-related documents/information reviewed and identified (i.e. review documents), interviews conducted, conclusions and findings, and recommendations.

The scope of the IR is set out in the Terms of Reference (ToR) which is publicly available (see Section 1.2.2). In summary, the IR's scope was on the project development and procurement process requiring consideration of three focus areas (and specific matters). The scope excluded project implementation and value for money matters.

We understand one of the main reasons for establishing the IR was in response to concerns about TG's delays and increased costs. To answer that question, we have identified four primary causes:

• A consented, non-PPP scheme design was used to procure a PPP project. This meant, in our view, that it was inevitable cost changes would occur as the TGP was better defined and an inputs-focussed design was adapted through PPP procurement to focus more on service needs and performance standards;

• The AT was set too low. The RFP Respondents realised this early in the RFP phase which saw Waka Kotahi revise the PSC and AT during the procurement phase. Also, and importantly, the decision to set the AT too low resulted in bid-responses that were value managed down (e.g. on design and timelines, as the main examples) to fit the AT number. These value-management changes represented risks that would eventually manifest themselves after procurement once construction commenced;

• Consenting risk could have been better managed by all the parties involved. In terms of Waka Kotahi's role, there was no formal/documented consenting strategy, the RFP issued to RFP Respondents was based on a Board of Inquiry (BOI) consents for a non-PPP project, which meant many of the 355 designation and resource consent conditions covered in the issued RFP would have required revisiting with consequential design, scheduling, and cost impacts; and

• From the review documents and interviews, we saw information that indicated that the TGP project governance could have been better, particularly the transparency as to how key PPP decisions were made. As we have raised already above, the main example we saw was that it appears the Project Team made the decision to set the AT - we expected to see stronger checks and balances when making key PPP decisions, particularly the clear articulation of the consequential impacts from setting the AT too low.

Note, we did identify that Waka Kotahi did initiate several design and safety changes during the procurement phase. We categorised these design changes as secondary causes to the increase in costs. Our reasoning for this view is that changes to the PSC post the Detailed Business Case (DBC) phase is a normal part of the PPP development process (and for non-PPP projects as well) as more detailed information becomes available.

Having made the above comments, we note that Waka Kotahi has embraced positively and diligently the need to ensure lessons learnt from the TGP are largely captured to improve future PPP road projects - this is happening with the Puhoi to Warkworth (P2W) project for which Waka Kotahi should be commended.

We also have another observation - the review facts show that, in a very short period of time, Waka Kotahi was required to prepare a PPP DBC using a consented scheme design developed for non-PPP delivery. If a "PPP-version" of the TGP had started around 2009, when the use of PPP was first signalled by the then Minister of Transport, then possibly the development of the TGP would have been more influenced by PPP requirements, potentially avoiding some of issues that the TGP has encountered.

Based on our review findings, we believe there are several changes that could be made by Waka Kotahi for future PPP projects, and also to aspects of New Zealand's PPP policy management. Table 1 (below) summarises our Review recommendations.

Table 1: Summary of Review Recmmendation

With reference to our Review's key focus areas, our high-level findings behind the above recommendations are summarised below (with detailed information provided in Sections 3, 4 and 5).

• Development of the PSC

Our view is that the budget presented in the 2012 PPP DBC was based on a reasonable supporting analysis and assessments of costs and risks, given the circumstances that the Project Team were facing at that time. Costs and risks were considered when developing the PSC which, using the definition available at the time to the Project Team, is the risk adjusted capital and operating costs of the project if it were to be delivered by Waka Kotahi, NZ Transport Agency using conventional procurement methods.

We noted from the review documents that Waka Kotahi were asked during a very short period of time (i.e. between August and September 2012) to convert what effectively was a non-PPP project, with an existing consented scheme design, into an initial PSC estimate that could be used to inform the DBC to confirm PPP delivery for the TGP. This situation meant Waka Kotahi had to use a design that was less developed than what they would normally include in a DBC.

Further, Waka Kotahi was also using project information that was not framed around a "PPP-lens" - for example, it is usual to see the outline of a Service Need Specification in a PPP detailed business case. Not having an outline Service Need Specification means that cost changes inevitably occur during the later procurement phase as a Service Need Specification is prepared, to inform the tender documents to which bidders are required to respond. A Service Need Specification is a vital element in PPP project's development, as it sets outcomes and services to be delivered, and performance standards to be met, which together influence whole-of-life project costs. As we raise later under Consenting Risk, the facts show TGP prior to August 2012 was developed on a non-PPP basis which means important PPP definition work (such as the Service Need Specification) were not done when the consented scheme design was developed (which was subsequently used to inform the PSC estimate in the DBC).

The one area of difference in the business case's costing methodology that we identified was for Operating and Maintenance (O&M) costs. The review documents explained that Waka Kotahi assumed a network/marginal costing approach to recurrent O&M costs - this is fundamentally different to the approach normally taken by a PPP consortium who would focus on specific recurrent costs and performance-related risks needed to maintain the TGP section as a standalone project from the rest of the road network.

We could not identify from the review documents if any PPP-benchmarking of O&M costs and estimates of transferred risk had been done for the DBC. We believe that Waka Kotahi should consider developing a cost and risk benchmarking database to assist managing their portfolio of road availability PPP projects. Note, we have also commented further under Other Matters about benchmarking once Transmission Gully (TG) commences the services phase after September 2021 (i.e. the revised date for construction completion).

The review documents showed Waka Kotahi advised the Waka Kotahi Board (Board) in December 2012 (after the PPP DBC was approved) of a planned review of the PSC in the first quarter of 2013, and that it was likely the PSC would change as the TGP was better defined as more information became available. A second review of the PSC was needed in August 2013 mid-way during the Request for Proposals (RFP) phase. This second review was in response to RFP Respondents' feedback that they were having difficulty meeting the AT figure set by Waka Kotahi. The main reasons for these second-round changes appear to centre on:

• Waka Kotahi-initiated changes to design requirements that were not in the consented scheme costing, mainly around an increase in the design speed to 110 km/hr (from the 100 km/hr design speed in the consented scheme estimate), road safety requirements, and revised structural requirements for bridges; and

• More detailed information coming to light, particularly as both RFP Respondents conducted their own due diligence on the seismic and geotechnical conditions to inform their technical solutions, as well as their view on how they would need to manage their consent obligations.

Our advice is that post-business case reviews of the PSC are common practice in PPP procurements. Our view is the post-business case reviews of the PSC were structured in-nature and would normally occur during a project's development (and also possibly if TG was implemented using non-PPP delivery methods). For example, the review documents show a total of $151.5 million in changes to the PSC from the two cost reviews with the main changes covering:

• Changes to inflation estimates ($32 million);

• Waka Kotahi-initiated scope changes, including changes to road design standards ($69.2 million); and

• Changes to risk allocations including road safety, seismic and geotechnical ($21.3 million).

Our main observation, though, is that the second review of the PSC in mid-Aug 2013 was triggered by concerns raised by the RFP Respondents about meeting the AT. Potentially, some of these cost impacts could have been incorporated into the first cost review of the PSC if they had been considered in more detail.

For example, from the review documents we identified the changes due to the design speed were being flagged by the Project Team in the weeks leading up to the release of the RFP in May 2013. From review documents, it appears that the revised design changes required confirmation in order to be included in the RFP documents, or potentially delay the issue of the RFP documents. The review documents record a decision made by the Project Team to include a revised design speed of 110 km/hr in the RFP documents and that the PSC risk allocation was sufficient to cover this requirement. The review facts, however, showed that the PSC costs needed to increase to allow for the greater design speed.

Our other key observation is that the record of that decision made by the Project Team regarding the cost impacts of changes to the design speed could have been supported by more detailed information. Further, such an important PPP decision should also have been endorsed by a higher-level governance group above the Project Team.

• Commercial case for the selection of the PPP model

Our overall opinion is that the commercial case used to justify the PPP model in the DBC to deliver the TGP was based on internationally accepted approaches and was informed through a structured decision-making process using valid qualitative and quantitative PPP assessment criteria.

Our only observations regarding the PPP commercial case relate to what we think could be clearer explanations for stakeholder groups on the following aspects of using the PPP model:

- The primary purpose of PPP business case is to decide if a PPP procurement is worth doing - a PPP business case does not, however, guarantee "cost certainty". It is usual during the procurement of PPP projects that costs will change as the project is better defined. We note the latter point can also apply to non-PPP delivery of major and complex infrastructure projects;

- The strategic policy/funding considerations about how to accelerate delivery of nationally significant infrastructure (such as the TGP) sometimes involves governments considering more borrowings, and that the PPP model represented a structured way of linking increased borrowing to better quality infrastructure and performance standards; and

- Using the PPP delivery model does not mean the financing impact for the commercial case is based on 100 percent private financing. If under non-PPP delivery the government would have financed a project through government borrowings, then technically the cost of private sector finance using PPP delivery is the difference between the risk-adjusted private and public sector borrowing rates.

• Whether the AT was sufficient, and how it was set and approved

Based on the review documents and interviews conducted, we conclude that that the setting of the AT was done using a lower P-value of the PSC approved by the Board as the borrowing/funding limit for TG, with the aim of driving the RFP Respondents' behaviour in the achievement of innovation and efficient whole of life solutions. (Note, in simple terms a "P-value" is a probability-based cost estimate that assumes a project's total cost will not be exceeded - i.e. an "upper limit").

This "innovation and efficiency" principle may be double counting insofar as a definition of the PSC should represent the "efficient" method of the government delivering the PPP project to the same specification and performance expected by the private sector. Consequently, the TG Project Team may not have fully appreciated that using the risk- adjusted PSC would have served their intended purpose of setting an "innovation and efficiency" target. As a result, the TG Project Team effectively set two "efficiency" targets (i.e. a risk-adjusted AT and risk adjusted PSC).

We think the Project Team may have been genuinely working to put a definition in place for the AT that was not available in the New Zealand PPP Policy at the time. However, we did note in the review documents another Project Team view that setting an AT too high created a risk of "gold plated" responses from the PPP market.

In terms of how the AT was approved, the review documents show that the Project Team made the final decision on where the AT was set. Whether any one member of the PT took the lead in this is unclear. We could find no evidence that this decision was either recommended to or endorsed by any higher level governance group. The Board appears only to have been advised that the AT was set using a (lower) P-value of the PSC to drive additional value.

Our view is the impact of setting the AT too low had major implications for the TGP - that is, RFP Respondents would have thought to be successful they had to "chase the AT". RFP Respondents were driven to submit a price set by the AT by value managing design and timelines in their bids (as main examples). We believe the risk and financial implications of making these value management changes would be reflected later in the construction phase when cost and timing pressures became evident.

Note, the review documents showed that the Board were advised by Waka Kotahi of the lesson learnt feedback that the tight AT had a direct impact on both commercial and technical aspects of bids, but it was considered that a tight AT is part of driving a good value solution and that these impacts are part of the process. Our view is the AT-related impacts and the confirmation that they were part of the process, as advised to the Board, did not appear to be the impacts that actually transpired.

• Whether the consenting strategy was appropriate, including the allocation of consenting risk and the impact Consenting Authorities may have had on project outcomes

Our observations on this specific matter based on the review documents and interviews are:

• The DBC only contains a one-page statement outlining the TGP's "consenting strategy". We did not see a formal project artefact that we would class as a management strategy for consenting risk, which also could have been used to provide more information to the RFP Respondents on what they needed to take into consideration when developing their RFP responses regarding managing consents;

• The TGP Agreement risk allocation matrix reflected a contractual strategy that all consenting risk is transferred to the PPP Consortium. This contractual strategy appears fine in-principle, but it is an approach structured on a risk allocation of BOI2 approved consents rather than the different requirements that were set as part of the RFP documentation and which came to light during the Preferred Bidder phase (and eventually into the construction phase);

• In effect, we believe the RFP issued could be considered to be an "unconsented RFP", given the number of changes that were required to be made. The review documents we saw listed a total of 355 designation and resource consent conditions requiring attention, and these conditions would all have needed to be factored into the detailed design for TG and preparation of outline (management) plans for implementing consents. Added to these 355 conditions were the subsequent consents, and changes to conditions from the local/regional government required to implement the BOI approvals and other conditions. Working through all these issues would have time and cost implications for TG;

• Waka Kotahi was aware during the procurement phase of the need for the Preferred Bidder to put more effort and resources into managing their consent obligations and took active steps to work with the Preferred Bidder to work on this aspect of their performance. However, this effort was being focussed after a Preferred Bidder had been selected when all parties would have been focussed on achieving Financial Close - obviously and in hindsight, providing more consenting-risk information and assistance to RFP Respondents earlier in the procurement process would have been a better approach; and

• The impact of consenting authorities is perhaps better explained as the responsibility left to the five local/regional authorities to manage the implementation of the BOI-approved consents. The review documents and interviews suggest these parties were genuinely trying to implement the intent of the BOI conditions, but there may not have been a common understanding amongst all project participants about how this was to be done in a coordinated and collaborative manner. Added to this the obviously challenging situation is that the local/regional authorities were also themselves coming up to speed with the availability PPP model and a major road infrastructure project.

Our overall observation from the review documents and interviews is that the TGP's consenting strategy clearly could have been better managed by all parties involved. For example, this includes both improving information to the market about their consenting obligations, but also using a "flexible" consenting strategy that aligns the consents with an outcomes-focussed procurement rather than being directly linked to a particular road design. These and other consenting management issues was reflected in lessons learnt information provided in the P2W DBC about how consenting would be done differently for P2W compared to the TGP. We specifically note Waka Kotahi has actively worked to implement many of these lessons learnt to ensure consenting risk is better managed for future PPP projects.

• Whether sufficient time was allowed for approvals, commencement and completion of the tender process

Our opinion is that generally the procurement timelines for TG were in line with international examples for similar road and availability PPPs. None of the interviewees highlighted major concerns to us regarding this focus area. The procurement timelines appeared to us to be benchmarked to example PPP procurements, and took into consideration that a drawn-out PPP tender has cost implications for the private sector participants.

Our view is the procurement timetable outcome was a major achievement given TG: (i) was the first road availability PPP by the New Zealand Government and for Waka Kotahi as an organisation, and (ii) is a highly complex infrastructure project and transaction.

• Whether the project governance structure had the knowledge and experience of delivering PPP construction projects

From the review documents and interviews, we saw good work by Waka Kotahi to mobilise quickly after the November 2012 decision to proceed with a PPP to procure experienced resources, leverage existing New Zealand capability and experience in PPPs, and set up an initial project governance structure. Waka Kotahi's ability to mobilise quickly and set up a project team management structure reflects its extensive capital works delivery experience.

We believe that the governance structure used for the TG PPP was an amalgam of how Waka Kotahi would traditionally deliver a major infrastructure project supplemented by PPP-specific advice. While we can appreciate that Waka Kotahi put in place a governance model that could have the best interfaces/alignment with the rest of the Waka Kotahi business, we think for a first availability PPP project that Waka Kotahi should have relied more on a benchmark PPP governance structure.

Our view is that the TGP governance structure may have inadvertently led to some expert advice not being heard at some of the right forums, or at the right time. This result can happen if (as reported in the Lessons Learnt reports about feedback on the TG procurement) a governance structure is perceived as complex, with multiple groups involved, and the respective roles/distinction between them is not clear. From the review material, it appeared to us the project governance structure could have had much clearer lines of distinction between governance/decision making versus advisory/management functions. The lack of such clarity not only results in a perception of a complex governance structure, but the real problem then becomes that the governance could possibly see too little of what they should be seeing - not just details but whole areas of decisions. Two examples of the latter we raise in this Report include setting the AT and sufficiency of risk allocations in the PSC to meet the increased cost associated with design changes.

To us, this suggests a governance structure that could have been tighter, particularly around who was making and approving key PPP recommendations. Our main example was the responsibility given to the Project Team to set the AT when the TG PPP decision making matrix assigned that responsibility, in the first instance, to a governance body above the Project Team. Even if the setting of the AT was "assigned" to the Project Team, this was not, in our view, clearly reported to the Board - this means a basic governance control of separating the setting of recommendations and approval/endorsement was not followed.

Another example we identified is the insufficient level of detail relating to the decision made by the Project Team to include a revised design speed of 110 km/hr in the RFP documents and that the PSC risk allocation was sufficient to cover this requirement. The review facts, however, showed that the PSC costs needed to increase to allow for the greater design speed.

Based on the above findings, we believe, for future PPPs, project governance and reporting to the Board should be reviewed and improved. This should include the Board endorsing a recommended PPP project governance structure and delegation matrix setting out roles and responsibilities for key PPP decisions.

We also noted from the review documents that the Board did not appear to have its own independent PPP adviser but instead relied on briefings from Waka Kotahi management, senior Project Team representatives, and experienced Project Team advisers (when invited to Board meetings for key decision papers). We believe the Board should appoint their own specialist and experienced PPP adviser, who would provide the Board with an independent view, Importantly also, the role of an independent adviser should be to flag to the Board PPP matters that it should be interested in.

We also note that during the TGP procurement, the Board would have been considering important PPP issues at the same time as normal Board matters. Our suggestion is that the Board may want to consider if a board sub-committee should be given the charter to deal with detailed PPP matters and report to the Board when key decisions are required.

In additional to the IR scope, we have summarised in Section 6, areas of best practice that we observed together with several "Other Matters" that Te Waihanga and key stakeholders may wish to consider. Our key recommendations raised in Other Matters are:

Benchmarking of Services in the PPP Contract

• The Board seek a briefing on benchmarking of the TGP-contract services in the lead-up to the completion of the TGP construction in September 2021, including how the benchmarking provisions in the current TGP contract will provide assurance to the Board that value-for money is being demonstrated during the services phase of the TGP; and

• Te Waihanga considers whether PPP five-year benchmarking provisions should be included in PPP guidance material available for government agencies, including a "departures" policy in circumstances where these benchmarking provisions are not being considered.

Establishing a PPP Contract Management Framework

• Waka Kotahi prepare for the Board's consideration an implementation plan for establishing a PPP contract management framework and function to manage and monitor the services-phase of the TGP. The scope of this implementation plan should include both the TGP-specific contract obligations and also integration dependencies with network operations and management.

Use of the AT in the Value-for-Money Evaluation

• Te Waihanga considers whether future TGP reviews should also include lessons learnt from the use of the AT in the value-for-money evaluation.

We would like to thank all interviewees for their participation and assistance, and Te Waihanga for providing project support to deliver this IR. Once TG is opened for use, we expect users will be commenting on the improvement in the "travel-experience". TG has been a difficult, first availability PPP for Waka Kotahi to deliver but the learnings have been largely captured and are being used positively to improve PPP capability and knowledge in their subsequent projects. Also, good and capable people have been involved in helping to get the TGP delivered - that commitment has also come from the PPP participants. The key challenge we now see for Waka Kotahi is making sure, post construction, that they have a focus on long-term contract management that is about how value-for-money and performance are to be monitored and measured for commercial outcomes over the long-term tenure of the TGP infrastructure investment.

__________________________________________________________________________

1 See Joint Ministers Media Statement (released 21 August 2020) - https://infracom.govt.nz/assets/Uploads/Review-into-Transmission-Gully-announced.pdf.

2 In New Zealand, nationally significant developments seeking approval under the Resource Management Act 1991 (RMA) may be assessed and approved by a Board of Inquiry. The RMA is the key legislation governing the assessment and approval of development in New Zealand, including major projects.