3.2.5  Asset Allocation

The government’s choice of asset allocation is interesting for two reasons. First, it determines the expected return from assets set aside to fund future primary deficits. Other things being equal, a higher portfolio return will lower the tax rate required to finance future government expenditure (and decrease the size of the asset pool required).

Second, the asset allocation will affect the stochastic properties of the tax rate since different portfolio compositions imply different volatilities. Importantly, a more volatile portfolio (for example, one heavily weighted in equities) need not necessarily result in a more volatile tax rate because tax rate volatility is dependent on the correlation between portfolio returns and the primary balance. In the spirit of Bohn (1990), there may be efficiency benefits from diversification of the government’s portfolio.

We model two alternative portfolios (All New Zealand Bond and All Foreign Equity) and compare the results with those of the base case.[14]

Source: Authors’ estimates

According to these results, a strategy of tax smoothing that involved debt repayment (and then bond accumulation) would be dominated by a balanced budget strategy. However, this ignores the fact that this strategy generates an asset portfolio of 118% of GDP by 2051/52 that is expected to finance another fifty years of tax smoothing.[15] As discussed earlier, the extent of the understatement of benefits is likely to considerably raise all of the savings figures in Table 6.

Focussing on the comparative results, note that the deadweight loss savings are quite sensitive to portfolio composition. On the basis of expected savings alone, the modelling suggests that the Crown should follow a particularly aggressive investment strategy. However, the volatility of the investment returns should also be considered.

Figure 6 compares the distribution of deadweight loss savings for different portfolio allocations. As expected, the width of the distribution for the foreign equities portfolio is considerably wider than the other portfolios. This reflects the volatile nature of equity returns. However, most of the risk appears to be on the upside since the whole distribution sits much higher up the scale than the other portfolios. Table 7 contains the probabilities that the total deadweight loss is lower than balanced budget for each of the portfolios.

Figure 6 – Distribution of deadweight loss savings over balanced budget

Source: Authors’ estimates

These probabilities suggest that, given our modelling assumptions, only a government with a very low risk tolerance could justify moving away from a strategy of investing all primary surpluses in foreign equities.[16] However, this result would be sensitive to our assumption about the size of the equity risk premium and the correlation coefficients.

Source: Authors’ estimates

Another way of thinking about risk is to consider the volatility of tax rates under various portfolio allocations. Figure 7 shows the year-to-year volatility in tax rates for different portfolio allocations. Throughout the period, the three tax smoothing investment strategies generate significantly lower year-to-year volatility in tax rates compared with the balanced budget strategy.[17] As the stock of assets accumulates under the tax smoothing strategies, return volatility contributes more to volatility in the operating balance, leading to an increase in tax rate volatility over time. This is particularly the case for the All Foreign Equity scenario, where tax rate volatility trends upwards.[18] This upward trend could be stemmed by dynamically adjusting the composition of the portfolio as it increases in size. For example, investing primary surpluses entirely in foreign equities until 2025/26, and then shifting to a balanced portfolio over a ten-year transition period, stifles the growth in year-to-year tax rate volatility in relation to the foreign equity strategy (see “Dynamic Port” scenario in Figure 7).[19]

Another consideration is that, by the end of the projection period, the median tax smoothing tax rates are considerably below the balanced budget tax rate, particularly for the foreign equity strategy. This raises the question of whether it is better to have a less volatile but higher tax rate or a more volatile but lower median tax rate. To illuminate this trade-off, we consider the probability that the All Foreign Equity tax rate is lower than the All New Zealand Bond and Balanced portfolio tax rates.

Figure 7 – Volatility of individual tax rate changes for various portfolios

Source: Authors’ estimates

Figure 8 – Probability the All Foreign Equity tax rate exceeds that of alternative strategies

Source: Authors’ estimates

Figure 8 shows that this outcome is almost certain for the next sixteen years. After that, the volatility of the large and growing foreign equity portfolio leads to some adverse outcomes. Consequently, the probability that the All Foreign Equity tax rate will be lower falls to 65.3% for the All New Zealand Bond portfolio and 58.0% for the Balanced portfolio. Adopting the dynamic allocation strategy lowers that probability even further by the end of the period. These results highlight the potential for significant one-off adverse outcomes associated with the All Foreign Equity strategy, and convey the sensitivity of the tax smoothing calculation to initial conditions.