4.3 Considerations when setting debt targets

There is no simple rule for governments to follow to determine a prudent level of government debt, or for setting upper and lower bounds to recognise the uncertain environment in which fiscal policy is operated. But there are several issues that should be brought to bear on decisions about a debt target. These include the size of the fiscal buffer needed to respond to economic shocks; the implications of government debt for the risk premium on borrowing; and the role of debt in funding capital expenditure. Debt targets should also take into account expected future spending pressures, such as those resulting from population ageing, which may warrant the accumulation of financial assets or a reduction of net debt to pre-fund some of these expected expenditure increases.

Fiscal buffers allow governments to continue to fund expenditure programmes during economic downturns through borrowing and/or selling financial assets rather than by increasing taxes. This allows "tax-smoothing" over time. Fiscal buffers are also beneficial from a fiscal stabilisation point-of-view as they avoid the need for governments to raise taxes to fund expenditure during economic downturns.

Consideration of the appropriate fiscal buffer should also take into account the wider set of vulnerabilities facing the economy, including contingent liabilities, such as explicit or implicit guarantees of the financial sector, the risk and likely impact of natural disaster events, exposure to terms of trade fluctuations, as well as the level of private sector debt. A number of these vulnerabilities, and particularly New Zealand's high level of external indebtedness, has been mentioned as a weakness in recent sovereign credit rating assessments of New Zealand. This is reflected in the following excerpts:

"NZ's household saving rate, though improving, remains negative. NZ's national savings/investment imbalance is the key structural weakness". (Fitch, 2011)

"... These strengths are moderated by New Zealand's very high external imbalances, which are accompanied by high household and agriculture sector debt, dependence on commodity income, and emerging fiscal pressures associated with its ageing population". (Standard and Poor's, 2011)

The liquidity of the government's balance sheet is also an important factor, as this also influences the ability of the government's finances to withstand shocks. A more liquid balance sheet improves the ability of the government to meet liquidity demands if access to credit markets is restricted.

Other factors that are often cited as relevant for the riskiness of government borrowing, including the ability of the government to repay and refinance debt, are the currency the government borrows in, the term of the borrowing, and who holds the debt. The vast majority of New Zealand government debt is denominated in New Zealand dollars. This means that if the New Zealand economy is hit by an adverse shock and the New Zealand dollar depreciates the government is required to repay maturing debt and/or interest payments in New Zealand dollars rather than another currency (which would equate to a higher New Zealand dollar amount given the depreciation of the currency).

New Zealand government debt is predominately in the form of government bonds, rather than Treasury bills. Government bonds have a maturity of greater than a year and pay fixed interest payments to the investor. Treasury bills have a maturity of less than a year and rather than paying a fixed interest payment to the investor, provide a return to the investor through an appreciation in the price of the bill. As at 30 September 2012, the government had approximately $63 billion of government bonds on issue, and $8 billion of Treasury bills. The New Zealand Debt Management Office (NZDMO) provides data on the maturity profile of outstanding debt. The average maturity of government bonds was around five and a half years, and for Treasury bills was around six months (NZDMO, 2012b).

Approximately 67% of government bonds are held by non-residents and 33% domestically (as at 31 May 2013). The proportion of debt held by non-residents is above average compared to other OECD countries but less than Australia. Given the tendency of "home bias" in investment decisions, refinancing risk is generally considered to be lower the more government debt is held domestically. The proportion of government debt held domestically has increased in recent years. This has been attributed, in part, to a greater demand for government bonds from commercial banks as a result of the Reserve Bank of New Zealand's prudential requirements, which require commercial banks to hold a greater proportion of liquid assets, such as government securities (see Hoskin, et. al., 2009).

Ireland provides an example of how rapidly government debt can increase as the result of economic and financial shocks. General net government debt in Ireland increased from around 11% of GDP in 2007 to over 100% of GDP in 2012. This is shown in Figure 4. This case study not only provides lessons about government fiscal management, but also broader policy settings, such as the monetary and exchange rate policy and prudential policy, and how the government controls contingent and implicit liabilities.

Figure 4: General government net debt, Ireland, 2006-2012 (% GDP)

Source: IMF (2012b) World Economic Outlook database, 2012

In the New Zealand context, the government introduced the Retail Deposit Guarantee Scheme (DGS) and Wholesale Funding Guarantee Facility (WFGF) for financial institutions in late 2008 in response to the international financial market turbulence. These schemes increased the contingent liabilities of the government. The New Zealand government had approximately $124 billion of guaranteed retail deposits and $6 billion of funding guaranteed through the WFGF in October 2009 (New Zealand Government, 2009). This compared to net debt at the time of around $17 billion. The DGS, and subsequent extension of the scheme have now ended, and the WFGF has around $3.4 billion of funding still under guarantee (as at 31 May 2013). Payments to depositors in failed financial institutions protected by the DGS totalled just over $2 billion and recoveries from those receiverships totalled just over $1 billion resulting in a cost the government around $1 billion. Participating entities in both schemes paid fees for the guarantees provided which totalled just under $0.5 billion.

Recent OECD and IMF work has highlighted the extent and implications of implicit guarantees of the financial sector for sovereign risk management and the functioning of financial markets (see for example, Schich et. al., 2011; and Ötker-Robe et. al., 2011). Internationally, changes to the regulation, supervision and resolution regimes for financial institutions are motivated by a desire to address some of these issues. In New Zealand the Reserve Bank of New Zealand (RBNZ) has introduced new liquidity requirements for banks and has proposed adopting the new Basel III capital requirements which are intended to increase the resilience of commercial banks to future financial shocks. The development of the Open Bank Resolution (OBR) tool has been motivated by a desire to provide an option other than liquidation or a government bailout to resolve a bank distress situation, where a private sector solution is not immediately available (RBNZ, 2012).

Fookes (2011) has attempted to assess the appropriate fiscal buffer for New Zealand by analysing the impact of fiscal and economic shocks on the government's fiscal position using shocks that have occurred in countries with similar characteristics to New Zealand (ie, countries with high external indebtedness and relatively low government debt). He examines the implications of two scenarios: an earthquake and an economic shock of the magnitude that hit Ireland and Spain during the Global Financial Crisis. The analysis concludes that compared with previous fiscal consolidations, the earthquake scenario is considered manageable, whereas the most severe scenario based on the economic and financial shock that hit Ireland over the 2008 to 2010 period is considered just manageable, assuming uninterrupted access to government debt funding markets. This work shows that having a starting level of net debt below 20% of GDP is an important condition for ensuring these shocks would be manageable.

An alternative methodology could be to take a stochastic approach by examining the probability and impact of a range of shocks to the fiscal position, based on historical information. This information could be used to examine the desirable level of government debt. As far as we are aware, such an approach has not been taken to examine the future stock of government debt in the New Zealand context, although stochastic approaches have been used to assess elements that impact on the long-term fiscal position, such as budget balance and population projections. For instance Buckle et. al. (2002) uses a structural vector autoregressive (SVAR) model to examine the impact of different shocks to the government's short-term budget balance. One of the insights from this paper is that even for short future time spans, projections of the budget balances in the New Zealand context involve a high degree of uncertainty which underscores the importance of a debt buffer. Stochastic approaches have also been applied to population projections in New Zealand (see Creedy and Scobie, 2002; Dunstan, 2011).

Stochastic approaches have not tended to be used in the Treasury's long-term fiscal modelling in the past because over long time periods there is so much variability around demographic, economic and fiscal variables that this approach would generate extremely large confidence intervals. The results in Buckle et. al. (2002) illustrate that even increasing the time horizon out from one to five years (let alone forty years) has a big impact on the confidence interval around the estimates. While the uncertainty of the future fiscal position is an important communications message, and may inform decisions about whether to act now, or delay and act later once more information is available, large probability distributions may provide a justification for inaction. The OECD (2009b) point out that while sensitivity analysis can be used to highlight the uncertainty that projections are subject to, too much sensitivity analysis can over-emphasise uncertainty and undermine the impact of projections (presumably encouraging delay in taking difficult decisions). Another consideration that will have implications for the timing of fiscal adjustment is the expected costs and benefits of bringing fiscal adjustment forward to reduce potential future fiscal costs or risks versus the benefits of additional information that can become available from delaying adjustment. This is discussed further in section 5 of the paper.

The potential for high levels of public debt to "crowd out" private investment is another factor to consider in setting a public debt target. Increasing government debt may be costly if large injections of debt-financed government spending crowd out private sector spending by driving up the real interest rate and real exchange rate. Hall et. al. (1998) develops a small open economy model which shows the impact of the fiscal balance on the interest rate premium. The scenario of a rise in the debt-to-GDP ratio of 2% over a two year period is found to be initially expansionary, but then neutral over the long term. The expansion comes from increased consumption and investment expenditure. The crowding out of private sector expenditure, including investment, is less than full. Key transmission mechanisms are the interest rate risk premia and real interest rates and an appreciation of the exchange rate. Baumol (1967) suggests that crowding out of private sector investment will be costly to the economy as a whole if the government is less productive than the private sector. High levels of government debt may also be costly for the economy if government debt pushes up the risk premium on borrowing for private individuals as well as the government.

Another consideration when setting debt targets is that governments may want to borrow to fund capital investment rather than fund investment through current taxation. Raising debt may allow governments to fund potentially growth-enhancing investments more efficiently than through raising taxes. This is the basis of the "golden rule of public finance" which states that the government will borrow only to invest and not to fund current spending. If specified over the economic cycle then this would mean that the government budget (excluding investment) must balance or be in surplus. This rule has been formalised in legislation in some countries. The golden rule was one of several fiscal principles set out in the United Kingdom's 1998 Public Finance Act. The rule was subsequently abandoned in 2009. The golden rule may be considered less transparent than a debt target, as it depends on where the economy is in the economic cycle. There are elements of the golden rule reflected in the New Zealand PFA, including that government needs to run small budget surpluses on average over time.

It may also be desirable to pre-fund government expenditure associated with population ageing. This may be justified on efficiency or on equity grounds. It may be more efficient for governments to pre-fund New Zealand Superannuation (NZS) for example, as long as the returns to investing in capital are higher than the growth in wages (see Coleman, 2011). It may also be more equitable to tax current generations to fund their future entitlements. However, in moving to more pre-funding there is always the issue of the transitional generation that needs to pay twice by continuing to fund entitlements of the current elderly, as well as pre-funding its own entitlements. This has both efficiency and equity implications. There are opportunity costs associated with pre-funding entitlements (eg, not being able to reduce taxes or increase expenditure in other areas now) and there may be governance costs associated with some pre-funding options (eg, establishing and running a superannuation fund).

Most of the above discussion has concerned setting upper limits for net or gross government debt. When considering whether there is a minimum level of gross debt that governments should hold, some considerations will include the benefits of maintaining a liquid market for government bonds, as well as the role of debt in funding capital expenditure, especially for long-lived assets (discussed above). Reinhart et. al. (2000) examines the economic implications of declining government debt in the United States. The paper suggests that if one of the reasons market participants buy US Treasury Bonds is because of their liquidity, reduced liquidity could result in the emergence of a new benchmark financial product with greater liquidity. As a result, the liquidity premium that market participants are willing to pay for Treasury bonds could be reduced. The Australian government established a "Future Fund" in 2006 to invest government budget surpluses, rather than using the surpluses to repay government debt. One of the motivations for this was that the government wanted to maintain a market for government bonds (Emmerson et. al., 2006).

In terms of the path to get to the chosen debt target the implications of fiscal adjustment on short-term growth will also need to be taken into account. Research into the growth effects of components of government expenditure and taxes suggest that the growth effects of changes in fiscal policy vary by the types of taxes and types of government spending. If for instance, the government had to reduce some investment due to it breaching its debt targets then that could, depending on the quality of the investment, have adverse long-run growth implications (see for instance Kneller et. al., 1999). Also, while New Zealand is thought to have small fiscal multipliers as a small open economy, fiscal consolidation is likely to have some short-term growth effects. It will also impact differently on different sectors of the economy.

Most of the above discussion has been about how government debt can be used as a mechanism to manage volatile revenues and provide a buffer to respond to economic shocks. However, there are other mechanisms, such as stabilisation funds, that can also be used in conjunction with debt to achieve these objectives. Stabilisation funds are designed to save temporary increases in government revenue (such as those from temporarily high terms of trade) in order to finance deficits in later years. For a discussion of stabilisation funds in the New Zealand context, see Brook (2013).