Part 2: Progress in repairing pipes and roads

2.1
In this Part, we discuss:

• the progress that has been made on repairing pipes and roads since 2013; and
• the Alliance Agreement's contribution to the horizontal infrastructure repairs.

Summary

2.2
Solid progress has been made on repairing pipes and roads (including retaining walls and bridges); by April 2016, 87% of SCIRT's repairs were complete.

2.3
The Alliance Agreement contributes to the effective and efficient rebuilding of the horizontal infrastructure assets. For example, the Alliance Agreement sets out how work should be allocated to delivery teams. Delivery teams now focus more on collaboration than competition for allocation of work. To date, the delivery teams have delivered work under budget.

Progress and changes since 2013

Solid progress has been made on repairing pipes and roads; by April 2016, 87% of SCIRT's repairs were complete. Since 2013, there have been changes in the staff and organisational structure of CERA and the Council.

2.4
In 2013, SCIRT had been established for two years. From May 2011 to October 2013, SCIRT had completed repairs to 200 km of wastewater pipes, 12 km of stormwater pipes, 42 km of freshwater pipes, and 279,576m2 of roads. By April 2016, 87% of SCIRT's horizontal infrastructure repairs were complete. Figure 2 shows the progress made by SCIRT by October 2013 and by April 2016.

Figure 2
Amount of repair work completed by SCIRT by October 2013 and by April 2016

Source: SCIRT.

2.5
Although the public entities involved have remained the same, there have been changes in staff and structure for both CERA and the Council. For example, at the local government elections in October 2013, a new Mayor and nine new councillors were elected. Also, the Council staff involved in horizontal infrastructure work underwent two restructurings. CERA was disestablished in April 2016, and its work distributed to other public entities. Some of CERA's role, including relating to horizontal infrastructure repairs, is carried out by a newly formed group in the Department of the Prime Minister and Cabinet.

The Alliance Agreement is mostly working as intended

The Alliance Agreement continues to contribute to the effective and efficient rebuilding of the horizontal infrastructure assets. Changes to the allocation of work to delivery teams means there is now more focus on collaboration than competition. Delivery teams are allocating more than the required minimum amount of work to subcontractors, and delivery teams have been delivering work under budget.

2.6
The Alliance Agreement encourages competition and collaboration between delivery teams in three ways: the method for allocating work to delivery teams (see paragraphs 2.7-2.13), allocating a minimum amount of work by cost to subcontractors (see paragraphs 2.14-2.16), and the "pain/gain" mechanism (see paragraphs 2.17-2.20).

Changes to allocation of work to delivery teams

2.7
Project allocation is part of the system of penalties and incentives in the Alliance Agreement to foster desirable behaviours and constrain cost inflation. Effective project allocation is important because it supports good performance in delivery teams.

2.8
The model for allocating work to delivery teams has two parts. Part A involves calculating a combined performance score for each delivery team. SCIRT calculates the score by assessing delivery teams' performance in five key result areas, as shown in Figure 3, to produce a delivery performance score. SCIRT then combines the delivery performance score with time and cost measures to form the combined performance score, which determines the target percentage of work to be allocated to each delivery team by cost (the target allocation).

2.9
In applying Part B of the model, SCIRT considers factors that might influence why a delivery team should not be allocated a project. These factors include the delivery team's available capability and capacity, proximity to a project, and safety performance.

Figure 3
Weighting of performance in key result areas for the delivery performance score, 2012 to 2015

* Because SCIRT considers safety as a separate factor in Part B, it has a weighting of 0% in Part A.
Source: SCIRT.

2.10
Some delivery teams are getting more work than their target allocation, while other delivery teams are getting less. The difference in the percentage between the target allocation and the actual allocation of work has fluctuated over time. Figure 4 shows the how much the actual allocation varied from the target allocation for each delivery team from July 2013 to February 2016.

Figure 4
Variation of actual allocation from the target allocation for each delivery team from July 2013 to February 2016

The line for each delivery team shows how much their actual allocation was either over or under their target allocation. A negative percentage means SCIRT is allocating less work to a delivery team than its target allocation. A positive percentage means SCIRT is allocating more work to a delivery team than its target allocation. For example, Delivery Team D was allocated 7.7% more work than its target allocation in May 2015.

Source: SCIRT. The letters assigned to the delivery teams for Figure 4 are different from those assigned for Figure 5.

2.11
The variation in the range is because the value of individual projects vary, making it difficult to allocate the exact amount of work by cost needed to give a delivery team its target allocation. Aligning the target and actual allocation also depends on funding being available, and the public entities approving individual projects for SCIRT to allocate to the delivery teams. Allocation was interrupted during the optimisation exercise (which reprioritised projects within the programme's remaining funds – see paragraph 5.16) and when the public entities put some projects on hold as a result of funding disagreements (see paragraphs 5.20-5.24).

2.12
In June 2015, the SCIRT Board approved a revision to the process of allocating work to delivery teams. The change was needed so SCIRT could complete the full programme of work before December 2016. The revision put more emphasis on Part B of the allocation process. This places weight on available capacity, capability, and the best outcomes for the programme rather than on delivery performance. It is too early to tell what effect the revision to the process of allocating work will have on the programme.

2.13
In our view, work has largely been allocated consistently with the Alliance Agreement. We consider the variations are understandable in the light of interruptions during the optimisation exercise and the funding disagreement, and the shift in focus needed to complete the work programme.

Work is allocated effectively to subcontractors

2.14
Under the Alliance Agreement, a minimum of 40% of the work completed by SCIRT, calculated by cost, must be subcontracted to parties outside SCIRT. The Alliance Agreement states that a competitive process must be followed to select subcontractors. The delivery teams are responsible for ensuring that subcontractors meet the same standards of operation and key result areas as they do.

2.15
Allocating a minimum amount of work to subcontractors helps to:

• maximise competition for work within SCIRT on the basis of price; and
• increase the capability and capacity of the subcontractors for the benefit of the construction industry.

2.16
As at September 2015, the delivery teams have been subcontracting 63.4% by overall cost of the work completed. Figure 5 shows the percentage of work subcontracted by each delivery team. All the delivery teams are subcontracting at least 40% of the work by cost. The lowest percentage is 48.6%, and the highest is 78.7%. This involvement of subcontractors is expected to increase skills and expertise in the construction industry, and to ensure competitive pricing of work.

Figure 5
Percentage of SCIRT work done by subcontractors, by cost and delivery team, as at September 2015

Source: SCIRT.
The letters assigned to the delivery teams for Figure 5 are different from those assigned for Figure 4.

Delivery teams deliver work under budget

2.17
The "pain/gain" mechanism in the Alliance Agreement is an incentive payment determined by financial and service performance. If the actual cost of a project is less than the target cost (the estimated cost of completing a project), a "gain" is created. If the actual cost is greater than the target cost, a "pain" is created.

2.18
The pain/gain payment is calculated at the end of the programme by adding up the pain/gain outcomes for every project. This is then shared 50/50 between the public entities and the delivery teams. The delivery teams' 50% share of the pain/gain is adjusted by the overall performance score (a score generated by the delivery teams' collective performance against the key result areas), and shared between each delivery team in proportion to the target cost value of the work they completed.

2.19
The purpose of the pain/gain mechanism is to motivate delivery teams to achieve the best cost results. The pain/gain mechanism is also meant to facilitate collaboration and encourage delivery teams to assist under-performing delivery teams to ensure that the final result will be a "gain", rather than a "pain".

2.20
The delivery teams have been steadily improving their performance in delivering work that is under budget. Figure 6 shows the estimated pain/gain position over time. A gain position is a negative amount because the actual cost is lower than the budgeted cost for all completed projects. The opposite also applies. Over time, there has been a shift from a "pain" position to a "gain" position.

Figure 6
Estimated pain/gain positions for the end of the programme based on work completed

Source: SCIRT.

SCIRT delivers value, but there are some concerns

2.21
Overall, the public entities told us they were getting good value from SCIRT. In our 2013 report, we found that, when relevant variables are considered, SCIRT projects seemed reasonably priced. After our 2013 report, the public entities commissioned a consultant and an engineering firm to review the process SCIRT uses to estimate the cost of each project. The review, completed in September 2014, found that "... the construction costs of the rebuild being undertaken by SCIRT are fair and reasonable and delivering good value for money under the prevailing circumstances."

2.22
The public entities considered SCIRT (and its predecessor, the Infrastructure Rebuild Management Office⁴) was an appropriate model for the emergency and post-response phases. However, some interviewees felt it would have been helpful if SCIRT had done an assessment of whether SCIRT continued to be appropriate after these phases.

2.23
SCIRT considers that the work allocation process has worked well. However, SCIRT also considers that early in the programme, delivery teams focused on maximising their delivery performance score against the other delivery teams so SCIRT would allocate more work to them. This meant that the delivery teams were less collaborative. We were told that this might have been because delivery teams initially understood the competitive elements of the Alliance better than the collaborative elements.

2.24
In our view, the Alliance Agreement has mostly worked as intended. However, there are lessons to be learned in understanding how the different incentives interact and affect the behaviours of the delivery teams. See, for example, the situation described in paragraph 2.23.

2.25
In our 2013 report, we said that, as circumstances change, the public entities need to consider whether the Alliance continues to be suitable. Each natural disaster will have its own complexities that will need to be taken into account when considering who will deliver the rebuild and how. In future responses to natural disasters, it would be worthwhile to consider the appropriateness of the delivery vehicle at each stage of the recovery.

4: Further information is available in Appendix 1 of our 2013 report, Effectiveness and efficiency of arrangements to repair pipes and roads in Christchurch, which is available on our website www.oag.govt.nz.