Appendix: Our audit report on the financial statements of the Government

The largest sources of revenue for the Government are income tax and goods and services tax. These revenue sources total $68.3 billion for the year ended 30 June 2017.

As outlined in note 2, income tax is subject to significant assumptions and judgements due to the timing differences between the reporting date and when taxpayers file tax returns.

In order to record tax revenue, judgement is applied to estimating:

• the amount of tax revenue to be collected from provisional taxpayers who have not yet filed their final tax return;
• the amount of tax revenue where payments have been received but no provisional or final tax return has been filed; and
• the amount of tax revenue to be collected from, or refunded to, taxpayers who are not subject to provisional tax.

I considered the recognition of tax revenue to be a key audit matter because of the significant assumptions and judgements applied.

We obtained an understanding of the systems, processes and controls in place over the receipt and review of provisional and final tax returns, tax assessments, and tax revenue receipts.

We assessed controls in place over significant reconciliation processes.

We tested the underlying data used in the various tax revenue estimation models to ensure that it was relevant and was used appropriately. This was performed by reviewing evidence to support key assumptions. The sensitivity of key assumptions was also tested.

We tested the reasonableness of the estimation models by checking actual revenue received related to previous financial years against estimates made in those years.

I am satisfied that the assumptions and judgements applied in estimating tax revenue are reasonable.

The Government owns significant physical assets totalling $144.6 billion.

The valuation of some of these assets requires significant judgement due to the uncertainties inherent in the valuation of these assets, the quality of data available and the benefits these assets provide. I have identified some specific assets where such judgements are evident.

As outlined in note 16, the state highway network has been valued at $23.8 billion at 30 June 2017 by an independent external valuer. The valuation is based on information from a number of databases that identify the asset components that make up the network (roads, bridges, culverts, etc.), and their expected useful lives. These asset components exclude land which is separately valued.

There are some uncertainties about the values assigned to different components (formation, bridges, etc.) of the state highway network due to limited information on quantities and useful lives within some databases and incomplete information relating to certain cost components.

Some of the costs associated with road construction (e.g. traffic management) in urban areas are assessed as being a significant part of the network that may potentially be undervalued. An allowance to recognise these costs has been included since 2014. However, such costs before then cannot be reliably measured and are currently excluded from the valuation.

We obtained an understanding of how the state highway network is valued. This involved confirming the competence, capabilities, and objectivity of the valuer, challenging the valuers’ key assumptions and assessing the valuation procedures, including the information extracted from databases.

We also carried out audit procedures to confirm that key controls were operating over the systems and processes used to record cost and other asset information related to the state highway network.

I am satisfied that the value of the state highway network at 30 June 2017 is reasonable and consistent with valuation practices, and that the disclosures outline the uncertainties about the valuation.

As outlined in note 16, the rail network has been valued at $939 million at 30 June 2017. In arriving at this value the freight and the metro transport parts of the network have been valued on different bases, reflecting the commercial nature of the freight part of the network and public benefit nature of the metro transport part of the network.

The extent to which the freight part of the network is commercial is open to debate. The Government is currently conducting a review of rail in New Zealand to consider KiwiRail’s operating structure, capital requirements and funding mechanisms. This review is expected to be completed by 30 June 2018, and is likely to result in changes that could affect the valuation of the freight part of the network.

As outlined in note 16, the valuation of the rail network could increase by up to $4.3 billion if the entire rail network was not considered commercial, and it was valued on an optimised depreciated replacement cost basis.

We considered the evidence around the commercial nature versus the public benefit nature of the freight part of the rail network. The evidence included reviewing:

• the State-owned Enterprises Act 1986;
• strategy documents;
• forecast results;
• correspondence setting out the Ministers’ expectations; and
• minutes from Board meetings.

As in past years, the evidence showed mixed results for the commercial nature versus the public benefit nature of the freight part of the rail network.

We also considered the terms of reference for the review of rail in New Zealand. The outcome of the current review will be key in deciding whether valuing the freight network on a commercial basis remains appropriate.

Due largely to the current review of rail, I am satisfied that the judgement to value the freight part of the network on a commercial basis for the current year, although marginal, is reasonable, and that the disclosures outline the significant judgements.

As outlined in note 16, the electricity generation assets, which are at least 51% owned by the Government, are valued at $15.9 billion at 30 June 2017.The valuation of these assets is carried out by specialist valuers because of the complexity and significance of the assumptions about the future prices of electricity, the generation costs, and the generation volumes that these assets will create.

As a result, small changes to these assumptions, in particular the forecast prices of electricity and the discount rates used to determine the present value of these prices, could significantly change the value of these assets.

We obtained an understanding of how electricity generation assets are valued. This involved confirming the competence, capabilities, and objectivity of the valuers, testing the valuers’ procedures for carrying out the valuations, including the information they used to carry them out, and challenging the valuers’ critical assumptions and judgements. We also used our own valuation specialists to assess the valuers’ procedures.

We tested the sensitivity of the key underlying assumptions used by the valuers to ensure that they were reasonable, and we compared the forecast prices of electricity to the expected longer-term wholesale prices and market data where it was available.

We also confirmed the underlying information held about assets by verifying asset purchases and disposals in the current period. This included testing whether there was adequate supporting documentation for those purchases and disposals. It also involved confirming the opening assets balances, and evaluating the related financial statement disclosures.

I am satisfied that the valuation of electricity generation assets at 30 June 2017 is reasonable, and the disclosures outline the sensitivity and the complexity of the valuation of electricity generation assets.

As outlined in note 16, the portfolio of social housing was revalued at 30 June 2017 at a fair value of $26.8 billion. The portfolio is valued on a “highest and best use” basis which is aligned to market prices for properties of a similar size and condition in the same geographical location.

As part of the Social Housing Reform Programme, the Government has been disposing of properties to other social housing providers. Prior to transferring properties to “held for sale”, the Government places encumbrances on the properties to ensure their purposes for social housing is maintained.

Encumbrances of this nature limit the highest and best use of the properties and mean their fair value is reduced.

The reduction in value has been calculated using a discounted cash flow model based on a series of assumptions about income and expenditure for the properties.

We assessed the competence, capability and objectivity of the valuer and challenged the valuation and approach adopted, which was consistent with prior years.

We considered the extent to which the valuation appropriately took account of a wide range of significant factors, which included:

• underlying movements in market prices in places where houses are located;
• property acquisitions through to balance date;
• sales experience during the year to 30 June 2017; and
• the market implications of concentrations of social housing.

We reviewed the integrity of the data provided to the valuer on which its valuation has been based and assessed the appropriateness of the adjustments made to the valuation as a result of any timing differences between the date of the valuation and 30 June 2017.

We assessed the appropriateness of the methodology and assumptions used to calculate the value of the properties held for sale that have encumbrances placed on them.

We engaged our own valuation experts to review the methodology and calculations to assess the appropriateness of the approach.

I am satisfied that the valuation of social housing, including properties held for sale, at 30 June 2017 is reasonable and that disclosures about the valuation are adequate.

As outlined in note 11, ACC’s outstanding claims liability has been valued at $37.7 billion at 30 June 2017 by an independent actuary.

Key assumptions used to value the outstanding claims liability include:

• selecting an appropriate risk-free discount rate to present value future cash flows;
• selecting an appropriate risk margin for the inherent uncertainty in the estimate of the present value of future cash flows;
• estimating the impact of inflation and innovation on future medical costs; and
• estimating the length of rehabilitation from injuries.

The sensitivity of each assumption is analysed in note 11. This sensitivity analysis indicates that assumptions are closely linked, cannot be viewed in isolation and changes in assumptions can have a large impact on the value of the liability as well as the actuarial gain or loss recognised.

We obtained an understanding of how ACC’s outstanding claims liability is valued by assessing the reasonableness of the approach taken to value the liability. We also reviewed the key assumptions adopted by ACC for each significant claim type to ensure these were appropriate.

We tested the systems, and controls and carried out detailed testing of the process for recording claims.

We tested key assumptions by evaluating them against past claims experience. We assessed the reasonableness of forecasts that diverged from past experience by looking at the evidence supporting the forecast.

We tested the reconciliations of the underlying claims data to ACC’s systems, examined the sensitivity analysis for movements in key assumptions, and evaluated the related financial statement disclosures.

I am satisfied that the assumptions and judgements applied in estimating ACC’s outstanding claims liability at 30 June 2017 are reasonable, and that disclosures outline the sensitivity of the valuation to changes in assumptions.

As outlined in note 20, the Government’s liability for public servants’ superannuation entitlements for past and current members under the Government Superannuation Fund has been valued at $11.0 billion at 30 June 2017 by an independent actuary.

The present value of the liability is sensitive to the estimated return on assets owned by the Fund, expected rates of salary increases for public servants who are members of the Fund and estimated inflation and discount rates.

The sensitivity of critical assumptions and judgements is analysed in note 20. This sensitivity analysis indicates that assumptions are closely linked, cannot be viewed in isolation and changes in assumptions can have a large impact on the value of the liability.

We obtained an understanding of how the Government’s liability for public servants’ superannuation entitlements is valued. This involved confirming the competence, capabilities, and objectivity of the actuary, as well as testing the actuary’s valuation procedures. We engaged our own valuation specialists to review the assumptions, judgements and procedures used to value the liability.

We tested key controls that ensure the completeness and accuracy of membership data, which was used in the actuary’s valuation.

We evaluated the appropriateness of key assumptions used in estimating the return on assets owned by the Fund and compared the expected rates of salary increases against external benchmarks.

I am satisfied that the Government’s liability for public servants’ superannuation entitlements at 30 June 2017 is reasonable, and that the disclosures outline the sensitivities of the valuation to changes in assumptions.

As outlined in note 26, as at 30 June 2017, the Government has financial assets of $133.4 billion, of which $75.3 billion are valued at fair value (and $58.1 billion are valued at amortised cost), and financial liabilities of $128.3 billion, of which $7.6 billion are valued at fair value (and $120.7 billion are valued at amortised cost).

Financial assets and liabilities measured at fair value include derivatives (which have a principal value of $232.5 billion), marketable securities, and share investments.

Where quoted market prices are not available to determine the value of financial assets and liabilities, fair value must be estimated. This is done by applying a valuation approach that is most appropriate for the asset or liability, such as using valuation models. Inputs into the models will use market data when available, otherwise inputs are derived from non-market data, which requires judgement.

The fair value of financial assets and financial liabilities that are valued using non-observable inputs are valued at $3.3 billion and $66 million respectively.

I considered the valuation of financial assets and liabilities to be a key audit matter because of the significant amounts involved and the judgements applied.

We obtained an understanding of the valuation techniques, controls and inputs used to determine the fair value of financial assets and liabilities.

We also carried out a range of audit procedures which reflected the nature of the financial assets and liabilities being valued, the valuation techniques adopted and the uncertainties that existed in determining their fair values. These audit procedures included:

• testing the internal controls in place over data relating to financial assets and liabilities that has been entered into financial and treasury systems;
• obtaining an understanding of the controls and valuation approaches applied, where a fund manager carries out the valuation;
• comparing the fair value of financial assets and liabilities to independent information and investigating any significant variances; and
• assessing the appropriateness of the inputs used for valuing financial assets and liabilities where the fair value was dependent on non-observable inputs.

I am satisfied that the fair values of financial assets and liabilities at 30 June 2017 are reasonable and that the disclosures outline the significant judgements.