The British statistician George E. P. Box is attributed as saying, “All models are wrong, but some are useful”. That astute observation goes to the crux of government’s Three Waters proposal. Is the modelling done by the Water Industry Commission for Scotland (WICS) useful?
If the WICS modelling is to be believed, there are substantial financial gains to be had from the amalgamation of local council water assets. It compares two scenarios: The status quo (remain as is) model, and the 3Waters (reform) model, which transfers council water infrastructure assets (and the debt liabilities directly relating to those assets) to one of four entities. In the case of Entity A (Auckland north) for example, it says under the 3Waters proposal the average water cost per household would be $800 by 2051 and $8,690 should local councils continue to provide water services.
There are two reasons why the figures coming out of the reform model look so good when compared to the status quo: The assumption regarding debt limits, and assumptions around efficiencies arising from amalgamation.
FarrierSwier explains the different debt assumptions used as follows:
Credit metric targets
Projected revenues were set to achieve given financial metrics over, or by the end of, the 30-year horizon
- For councils, a ratio of net debt to revenue of no more than 2.5x was targeted
- For amalgamated entities, 2 credit metrics were used:
– a ratio of funds from operations (FFO) to net debt of at least 10% was targeted
– a FFO interest coverage ratio of at least 2.5x was targeted
which align with a Moody’s credit rating at the lower end of Baa or higher end of Ba.
In other words, the 3Waters reform model removes the 2.5x debt to revenue cap and replaces it with “Cash shortfalls funded by debt” where debt is ‘capped’ to retain a Moody’s Ba credit rating. 
This is evident in spreadsheet modelling behind the forecast which shows the debt to revenue ratio rising to 6.2x over 30 years in the 3Waters scenario but remaining constant at 2.5x in the status quo scenario.
So how significant is this debt assumption?
Very significant according to Castalia consultants who have peer reviewed the government’s modelling. A report prepared for the Whangarei District Council  confirms that under the WICS model the debt would rise from 2.5x to 5.82x. This, they say, results in the cost payable per household falling by 49%. In other words, approximately half of the assumed “saving” from the amalgamation model is due to the entities increasing debt instead of funding it by charging water users.
The consequence of increased debt in the 3Waters model is increased risk. A credible model would make an adjustment for risk, or at least report on a risk-adjusted basis. That risk is visible in the underlying assumption that the 3Water entities can borrow at a fixed rate of 3.5% for the next 30 years. That rate is now unrealistic, as is evident from leaked briefing documents relating to a debt blow-out at Kāinga Ora which reported the average interest rate on debt is forecast to be 4.6%, up from the 3.3%.
It is also unrealistic to assume a fixed rate of 3.5% when 30-year term fixed rate loans are not available. The debt will invariably have a short-term floating rate component with hedging contracts to mitigate interest rate volatility. The cost of this hedging is not insignificant but has nevertheless been ignored.
In reality, the cost of the debt is likely to be substantially higher than 3.5%. That increase will either be passed onto water consumers or added to debt.
Opex and capex assumptions
The remaining (51%) efficiencies are assumed to arise from the 3Waters reform model – the most significant and the most sensitive being the savings they assume can be made in operating costs (opex) and capital spending (capex).
With respect to capital expenditure (the cost of installing new treatment plants, pipes, building dams, etc) Castalia say:
WICS claims that the Reform Scenario will result in 50 percent lower capital costs. WICS claims that Entity A will progressively improve its capex efficiency so that by 2041 it is saving 50 percent per annum. That is, by 2041, for each $0.50 invested, Entity A will get $1.00 of capex value. This is an implausible assumption…
In our view…significant care should be taken when relying on the capital efficiency gaps estimated by WICS. This is particularly important, given the significant step up in investment forecast for the 30-year period and the role that the capex efficiency assumption plays when estimating benefits from amalgamation and associated reform. 
With respect to operating expenditure (staff costs, contractors, overheads) Castalia say:
WICS assumes implausible opex savings. Globally the major operating costs for water services are labour, third party (that is outsourced) services and materials and energy. New Zealand is no different. WICS claims the mega entities achieve opex efficiencies of between 53.3 and 61.9 percent by 2040 derived from econometric studies of UK water entities. Opex efficiencies achieved in the UK water sector are not a reasonable guide to the efficiency gap in New Zealand. Opex efficiencies above 50 percent in under 20 years is not plausible in the New Zealand water sector.
They go on to say this is because the government has promised that not only will all staff in council organisations be retained but a further 6,000 to 9,000 jobs will be created from the reform (!). Castalia also say the outsource service provider market is already competitive.
A further factor not adequately addressed in the WICS model is the effect of the co-governance arrangements on opex and capex. That influence is enshrined in Sections 140 and 141 of the Water Services Entities Bill now before Parliament, which deals with the rights of iwi.
Section 140 states:
(1) Mana whenua whose rohe or takiwā includes a freshwater body in the service area of a water services entity may provide the entity with a Te Mana o te Wai statement for water services.
(2) A Te Mana o te Wai statement for water services provided under subsection (1) may—
(a) be provided by an individual iwi or hapū, or by a group of iwi or hapū:
(b) relate to 1 freshwater body, or to multiple freshwater bodies.
(3) Mana whenua who have provided a Te Mana o te Wai statement for water services under subsection (1)—
(a) may review the statement at any time; and
(b) following a review, may provide a new statement that replaces the statement that was reviewed, in which case, the reviewed statement expires when it is replaced.
Section 141 states,
(1) As soon as practicable after receiving a Te Mana o te Wai statement for water services under section 140, the board of a water services entity must—
(a) acknowledge receipt of the statement; and
(b) engage with the mana whenua who provided the statement in accordance with section 202 in relation to the preparation of a response to the Te Mana o te Wai statement for water services.
(2) A response to a Te Mana o te Wai statement for water services must include a plan that sets out how the water services entity intends (consistent with, and without limiting, section 4(1)(b)) to give effect to Te Mana o te Wai, to the extent that it applies to the entity’s duties, functions, and powers. [emphasis added]
These provisions give local iwi and hapu significant influence over the water entities. There will clearly be a cost to the water Entity to engage with iwi/hapu but there is also an unanswered question about the degree to which the Mana whenua clauses will impact capital spending priorities.
There is no benchmark to estimate the potential cost of the mana whenua clauses as no other country has embarked on a management structure where a minority interest group is given such influence. It is evident from other co-governance arrangements – like the DoC-funded visitor information centre at Punakaiki – that iwi act in the best interests of iwi.
A further issue not addressed in the WICS model is whether the complex governance arrangements will prevent the entities from having “incentivised management” that can “achieve high levels of business performance, comparable to privately owned entities”, which WICS say is required if the opex and capex efficiencies are to be achieved. It is quite possible, indeed likely, that the cultural considerations embedded within the governance and management structures will create inefficiencies that may negate the efficiency assumptions made by WICS.
Given these limitations, it is not credible to suggest the new water entities will be so efficient that they can do for less than 50 cents what councils do for $1, or that they can build infrastructure for half the price that it currently costs a local council.
There is no evidence that efficiencies have been achieved from any of the amalgamations the Ardern government has undertaken since assuming office in 2017. The evidence is that the efficiencies promised are not realised in practice.
So, what is the answer to the question posed by George E. P. Box? Is the WICS model useful?
The short answer is: No.
The Reform Scenario is based on faulty assumptions and flawed analysis. The government has not shown with sufficient certainty to WDC that the claimed benefits of the Reform Scenario will materialise. 
FarrierSwier say the,
WICS analysis cannot be used to definitively conclude that amalgamation in and of itself will lead to material efficiency gains in New Zealand” and that “significant care should be taken when relying on the capital efficiency gaps estimated by WICS. 
Despite the reservations from experts the government still maintains the model proves the case for change. It does not.
To put it plainly, the underlying assumptions made in the government’s modelling are unrealistic. This creates significant risks to water users and central government.
The most likely long-term outcome of the 3Waters reform is that the assumed cost and capital efficiencies will not materialise, due to a cumbersome governance structure subservient to cultural interests. WICS make it very clear that the efficiencies will only arise if the entities are run with private sector management practices and disciplines. That would seem impossible given the complex governance arrangements and in particular the direct influence of iwi/hapu.
When considered in the context of the risks associated with the high debt, one must conclude that it is likely that the water entities will not remain viable without central government support or without additional cost to water consumers.
The water reforms as proposed in 3Waters is a high-risk proposition, with much higher risks than that faced by the 67 local authorities individually.
 Three Waters Reform, FarrierSwier, 2 May 2021. Page 5.
 Moody’s describe a “Ba” rating as “speculative” and subject to “substantial credit risk”. https://www.moodys.com/sites/products/productattachments/ap075378_1_1408_ki.pdf
 FarrierSwier. Page 66.
 FarrierSwier. Page 29.