3.7  Application to policy

Results such as those just presented could be included in the suite of fiscal indicators that informs the government's fiscal strategy. In addition, the model could be used to analyse the effect on risk of changes in the composition of the Crown's balance sheet. We illustrate this with a hypothetical policy change, in which half the Debt Management Office's unmatched debt portfolio (about $23 billion as at 30 June 2009) is replaced with US dollar-denominated debt. Our purpose is to model and analyse sensitivity of risk to policy, not to propose or even seriously discuss possible policy changes. We recognise that policy decisions will be based on a much wider set of considerations than those discussed here. We emphasise that further refinement of parameter estimation would be necessary to produce robust results.

We estimate that the volatility of US-dollar debt is significantly higher than the volatility NZ-dollar debt, because of the exposure to exchange-rate fluctuations.^[17] Nevertheless, the change could reduce the Crown's exposure to risk if US-dollar debt is less correlated with other Crown liabilities or more correlated with the Crown's assets.

We find that issuing US-dollar debt would have ambiguous effects on the Crown's risk exposure. The model indicates that GAAP net worth and comprehensive net worth would be exposed to more risk, but net worth excluding social assets would be exposed to less risk. In the latter case, the effect of the correlations outweighs the effect of the higher variance of the US-dollar debt. To the extent that weight is attached to the net worth excluding social assets, the result casts some doubt on the conventional view that domestic currency debt is lower risk. It is possible that the conventional view attaches too much weight to the volatility of the value of foreign-currency debt and too little weight to its correlations with other assets and liabilities on the Crown's balance sheet. But it is also possible that our estimates of correlations are wrong. Or, even if they are accurate estimates of average correlations, it may be that foreign-currency debt is particularly risky during bad times. During the recent crisis, for example, the New Zealand dollar fell sharply against the US dollar, just when the present value of tax revenue also fell sharply.

Previous empirical work has also raised questions about optimal currency composition of debt. Fowlie and Wright (1997) used a tax-smoothing approach that found that there were benefits from issuing foreign currency debt over 1985 to 1995. However, other work, such as Hawkesby and Wright (1997), arrived at the opposite conclusion, finding that the issuance of nominal domestic debt was preferable because it hedged supply-side economic shocks.

Note: Figures shown represent the difference between the results for this scenario and the values reported in Table 7. Positive (negative) values denote an increase (decrease) in value-at-risk.