Appendix II: Qualitative Assessment of Risks

This Appendix applies the indicators developed in Section 3.3 to assess qualitatively the risk properties of alternative policy options. The Appendix is in two parts: Part A discusses the key characteristics of the distortionary tax, time-consistency and agency cost objectives in general terms without reference to specific loss functions. Part B applies these general comments to the loss functions relating to the eight policy options available .

Part A: General Discussion

1.  Distortionary taxation

The two possible decision errors in relation to distortionary taxation are:

• False positive error: False conclusion that tax smoothing would confer a significant benefit. A decision to target the minimum-variance portfolio (and possibly build a positive CNW buffer and place an upper bound on total gross debt) that in fact confer no significant benefit relative to the absence of such targets; and
• False negative error: False conclusion that tax smoothing is an insignificant issue. A decision to adopt no specific risk/return target (or other bounds) that in fact would confer a significant benefit had they been adopted.

Implementation risk

Implementation risk could be substantive under the case where tax smoothing is significant. In particular, uncertainty surrounds the risk/return properties of any particular asset or liability, including those on the Crown balance sheet.

Information revelation

Apart from implementation risks, the probability and timing that a false positive or false negative error would be revealed as a mistake is similar across the two cases.

Consider a false positive that tax smoothing would confer significant benefit. Although a stable tax rate would be observed over time, the supposed welfare gains are unobservable. In particular, the difficulties of distinguishing econometrically the tax effects from other influences on economic performance suggests that ex post analysis would be unlikely to yield substantively new and more powerful information than available in the literature currently. (The main information that would come available following adoption of a tax smoothing policy would be the difficulties or otherwise of successful implementation).

Similar comments apply in the case of a false negative that the benefits of tax smoothing would be insignificant.

Reversibility

Assuming a policy mistake did become known, both cases should be low cost to reverse:

• in the case where a false positive (in favour of tax smoothing) was revealed, the Crown would be faced with unwinding asset positions that had been built up specifically for smoothing purposes. Provided the unwinding is conducted in an orderly manner, the cost of reversal should be fairly low; and
• In the case where a false negative (against tax smoothing) was revealed, the government would have the option of establishing the tax-smoothing regime. The legislative and other institutional arrangements would take some time to work through but would be relatively low cost. Losses would be incurred to the extent that the delay resulted in some or all of the window of opportunity passing by (e.g. with population ageing profile over next few decades).

Worst scenario

Losses under the worst scenario in the case of a false positive (in favour of tax smoothing) may or may not be larger than in the case of a false negative (against tax smoothing).

Under a false positive the worst scenario would be a substantial permanent reduction in asset values accompanied by failure to hedge the Crown balance sheet due to instability in correlations between assets. Under a false negative the worst scenario would be the loss in economic performance of the country due to instability in the tax rate. The loss would be greater to the extent that, for example, population ageing is a one-off change and the delay in implementing tax-smoothing mean the window of opportunity passed by.

Dependence on state variables

The loss arising from a false positive (in favour of tax smoothing) would be less sensitive to state variables than the loss arising from a false negative (against tax smoothing):

• A false positive means, by assumption, either that variation in tax rates has insignificant impact on economic welfare or that implementation is too difficult and costly.
• A false negative means, by assumption, that variation in tax rates does have significant impact on economic welfare (over and above the cost of implementation). Therefore the cost of being wrong would depend on the state of the economy and the profile of government expenditure.

Vote misperception risk

A correct conclusion in favour of tax smoothing faces the problem that the public would tend to look at specific measurable outcomes that may not be closely related to the policy objective. For example, a desirable hedging strategy could lead to an outcome where the market value of Crown financial assets was revised downward by (say) $5 billion as an offset to upward revaluation in the tax asset. Because accounting rules mean that the highly visible headline Operating Balance would include the downward revision in financial asset value but exclude the increase in the tax asset, the public would likely view the outcome as reflecting poor economic management by the government of the day. The hedging benefits are difficult to communicate. In contrast, a correct conclusion against tax smoothing would not face the same issues.

A tax smoothing policy shifts the burden of taxation over time. The current situation for New Zealand is that tax smoothing implies accumulating assets in preparation for government expenses associated with population ageing and may also imply a need to build a positive CNW balance (if tax rates and consumption are negatively correlated). These imply bearing the cost of a higher tax rate now in return for lower tax rate (relative to counterfactual) in several decades in the future.