2.4  But central risk management is limited

In the late 1990s and early 2000s, the Treasury intensively investigated Crown financial policy. Huther (1998) and Fabling (2002) constructed comprehensive balance sheets and estimated the expected return and the variance of the Crown's portfolio. Davis (2002) estimated uncertainty in the Crown's long-term fiscal projections. But, at least until very recently, this work had slowed nearly to a standstill. And the measurement that the government does do—including the sensitivity and scenario analysis—falls short of an analysis of risk: it provides no estimate, even approximate, of the likelihood of various outcomes. Thus the government currently has no estimate of the Crown's exposure to risk and no way of estimating how that exposure varies with policy.

In the absence of quantitative estimates of the Crown's exposure to risk, ministers must make (and officials must recommend) decisions about risk management on the basis of intuition or partial models. Intuitions about risk are, however, known to be systematically flawed. Uncertainty, for example, is frequently underestimated (see, for example, Kahneman and Tversky, 1979). And models that examine only parts of the Crown's portfolio cannot incorporate all the relevant correlations and therefore cannot adequately address questions about the Crown's exposure to risk. Although intuition and partial models can be helpful—and no comprehensive model of risk will be good enough to supplant them—exclusive reliance on them is problematic.

The range of relevant risk-management questions is wide. Do current policies expose the Crown to too much risk? Should the deficit be reduced more quickly? Or are the risks small enough that more attention should be paid to continuing fiscal stimulus? Should the Crown's exposure to risk be considered when choosing whether to alter the composition of the tax base (for example increasing GST and reducing corporate tax) or the risk-related effects small enough to ignore? Do contributions to the NZSF have a sufficiently attractive profile of risk and expected return that they should be resumed as quickly as possible? Or, on the contrary, should the NZSF be liquidated and the proceeds used to repay debt? Does foreign-currency debt really offer an unattractive combination of cost and risk? Or would some net foreign-currency debt be better than none?

Consistent with the lack of measurement, the Crown holds assets and liabilities in several portfolios that are managed more or less independently by various different agencies (the Debt Management Office, the NZSF, the Earthquake Commission, and so on). Although these agencies manage assets and liabilities for a common owner, each agency manages its own portfolios without much consideration of the interrelationships between portfolios, and thus the effects of their choices on the Crown's exposure to risk. Nor can the agencies be sure what the Crown's risk-related objectives are. True, the agencies are governed by rules that encourage them to act in the interests of the Crown, but the rules are too general to ensure appropriate actions in relation to risk.

One of the goals of the NZSF, to take just one example, is ‘maximising return without undue risk to the Fund as a whole'. The risk to the Fund, however, may differ substantially from the effect of the Fund's choices on the risk to the Crown. To the extent that the NZSF's managers consider assets and liabilities that are not part of the Fund, they limit their attention to the liability (comprehensively defined) for which the Fund is named; ‘we plan', they say, ‘to ensure we maximise our contribution to future superannuation payments by closely aligning our assets with the characteristics that drive future superannuation payments' (Guardians of New Zealand Superannuation, 2008).

In the absence of central direction, the managers of components of the Crown's comprehensive portfolio may inadvertently take actions that are not in the Crown's interest. They might expose the Crown to too much risk or pay too much to reduce risks that, from the Crown's perspective, are very small: risks that appear large to managers of the components of the Crown's portfolios may be insignificant for the Crown.[4] Public-sector managers may be especially inclined to excessive risk aversion, given the publicity that attends bad outcomes. Worse, if the value of a particular portfolio is naturally negatively correlated with the Crown's comprehensive net worth, the manager of that portfolio may spend money to reduce the risk of the portfolio in a way that increases the Crown's exposure to risk.

Well-intentioned attempts to reduce the Crown's exposure may also fail for lack of coordination. As noted above, the NZSF has contemplated choosing its assets with a view to reducing the risk of a larger portfolio containing its assets and the government's (comprehensive) superannuation liabilities. The Debt Management Office has also contemplated matching the debt it issues against some of the NZSF's assets. If both agencies were to match their assets against the same set of liabilities, the result would be one unintended by either, and there would no reason to think that the Crown's exposure to risk had been reduced.

Lastly, the Crown may fail to take simple actions to reduce its exposure to risk at low cost. For example, the Crown may be exposed to more counterparty credit risk than is necessary. As part of their operations, several Crown agencies enter into derivative contracts with various banks. At any one time, some agencies may owe a given bank money, while others may be owed money by the bank. Unless the Crown enters into a master agreement, however, the Crown cannot net off contracts that are assets for the Crown against those that are liabilities for the Crown. If the bank fails, the agencies that owe money must pay, while the others must join the creditors' queue.[5]