2.3 Long-term fiscal model

As mentioned above, the New Zealand Treasury LTFM converts information from the government's accounts into forward-looking projectionsof what would happen under certain assumptions.[9] Longer-term projections provide insights into possible structural changes in government expenditure and revenue over time in light of demographic ageing and other pressures. As a baseline what we present in this paper are projections from the Treasury's long-term fiscal model. These projections assume that government expenditure grows according to historical growth pressures. We will refer to this throughout the paper as the cost pressure scenario.[10] This scenario could also be interpreted as a "current policy" scenario which tries to capture policy as reflected in legislation or the history of decisions governments have made.[11]

The main assumptions in the cost pressure scenario are[12]:

1. The economic projections begin after the end of the Treasury's five-year economic and fiscal forecasts.
2. It is assumed the government implements its announced strategy for future operating allowances (in other words, government spending other than that on welfare entitlements and interest costs) during its current electoral term (for the three years following the last General Election). That is, the cost pressure, or "bottom up" expenditure projections are assumed to begin after that point (from 2015/16 onwards).[13]
3. Economic output is determined by: population, labour force, and labour force participation projections (central projections from Statistics New Zealand), and an assumed number of average weekly hours worked (currently 33.2 hours per week), unemployment rate (4.5%), and economy-wide labour productivity growth (1.5% per annum).[14]
4. Price inflation is 2% per annum (the mid-point of the target range in the RBNZ Policy Targets Agreement).
5. Nominal interest rate on government debt is assumed to move up gradually from the present 2.2% to 5.5% in 2021/22, remain at this level until 2026/27, and then climb to 6% from 2031/32.
6. Nominal government expenditure (excluding transfer payments and debt servicing costs) grows with the sum of: inflation (assumed to be 2%), real input price growth (1.2%), public sector productivity growth (0.3%), demographic demand growth (growth in the relevant recipient population with cost weights applied), and non-demographic demand growth (based on historic trends in spending patterns, that might reflect the response of demand to higher incomes or the availability of new technologies).
7. Spending on welfare benefits is projected to grow by the sum of the projected growth in the recipient population (as determined by current policy) and the price growth as determined by the particular indexation regime (for example, nominal wage growth for New Zealand Superannuation and a mix of the CPI growth and nominal wage growth for working-age benefits).
8. Assets and non-debt liabilities of the government are assumed to grow in line with either nominal GDP, CPI-measured inflation or from generated tracks produced by satellite models such as the Government Superannuation Fund (GSF) Model.
9. Government revenue is maintained at around 29% of GDP after the end of the forecast and transition period and/or once the average personal tax rate reaches its historical level, and
10. Core Crown gross debt is the residual of the model.[15]

Whether a government debt target is met is, therefore, determined by the level and growth of economic output, government revenue and expenditure decisions, as well as the government's cost of borrowing. These are the components of the expressions derived in section 2.1. The Treasury LTFM can therefore be viewed as an elaboration of the finite horizon inter-temporal budget constraint, which allows one to understand how population dynamics and other important influences impinge on the government's budget constraint over a period of time (in contrast to summary measures such as the inter-temporal fiscal gap). For example, demographic change influences the level and growth of economic output by influencing the size and growth of the labour force and labour force participation rates. It also influences spending on age-related government programmes such as education, New Zealand Superannuation and health services.

The LTFM can, therefore, be used to assess the sensitivity of fiscal projections to assumptions about labour force participation, migration, and so on. The IMF (2012a) has noted the tendency for projections of life expectancy for different countries to underestimate life spans consistently. Scenarios can be run using the LTFM to show the sensitivity of the projections to different life expectancy assumptions. It can also be used to assess the sensitivity of long-term fiscal projections to initial conditions, such as the likelihood of current governments maintaining their short-term fiscal programme. The approach for the 2013 Treasury statement is to assume that the current government implements its strategy for operating allowance-controlled expenditures (ie, expenditure on health, education, justice, etc., excluding welfare entitlements and interest costs) for the three years following the 2011 general election and to begin the expenditure projections for operating allowance-controlled expenditures, based on "bottom up" cost pressures, after that point.

The choice of when to begin the cost-pressure projections in these spending areas can be crucial to the time-profile of government debt. For example, beginning the fiscal projections for operating allowance-controlled expenditure three years after the last general election (in 2015/16), rather than after the five-year forecast period (in 2017/18), results in net debt being projected to increase to 200% of GDP by 2060, rather than 125% of GDP. This reflects the effect of compounding interest costs on government debt over time.

Bell and Rodway (2011) examine the reasons why the Treasury's net debt projections changed between 2009 and 2011. The net debt projections done in 2009 reached 223% of GDP by 2050, while those done in 2010 reached 106% of GDP by 2050 and those done in 2011 reached 44% of GDP by 2050. This change is largely explained by changes in the forecasts used for the initial five years of the projection period as a result of stronger taxation revenue growth arising from the improved economic outlook, and policy-induced government spending reductions. This comparison shows how important the medium-term fiscal projections can be for the time profile of the longer-term debt projections.

The treatment of taxation as well as some expenditure components in the Treasury LTFM (and many other long-term fiscal projection models) is fairly stylised. For instance, the 2009 version of the Treasury LTFM assumed that the proportion of public health expenditure that goes to each age group in the past will continue into the future. So for example, a 65-year-old in 20-years' time is assumed to demand the same proportion of health expenditure as a 65-year-old today, even though an average 65-year-old in the future is expected to healthier on average than a 65-year-old today. In the 2013 Statement, the Treasury builds in assumptions of "healthy ageing" into its projections, where projected increases in life expectancy are assumed to delay the increases in public health expenditure accruing to different age groups.

Previous versions of Treasury's LTFM have assumed that tax revenue as a proportion of GDP is independent of the age structure of the population. This is consistent with the work of Creedy et al. (2010) that shows that population ageing is expected to influence the mix of taxation revenue more so than the level of taxation revenue.

As well as defining the path of fiscal aggregates based on the cost-pressure scenario, it is possible to derive the budgetary changes that would be required to stabilise debt at a particular level (such as a target debt level). This is referred to as the sustainable-debt scenario. A net debt limit of 20% of GDP is used in the 2013 Statement to show the fiscal adjustment that would be required to stabilise net debt at this ratio. The way government spending was restricted in this scenario is to set annual allowances for new operating spending that stabilise net debt at 20% of GDP and then allocate the operating allowances to the key spending areas, such as health and education. It is important to note that the Treasury's long-term fiscal statements do not recommend a particular long-term government debt target, as this will change over time for a number of reasons (this is discussed in the next section of this paper).

Scenarios can be run to show the sensitivity of the projections to different government debt targets, for example targets ranging from 0% to 30% of net government debt to GDP. Interestingly, within reasonable bounds the government debt target doesn't make a big difference to the fiscal adjustment that is required. This is illustrated by the fiscal gap calculations in section 4.7 of the paper. Other scenarios are of course possible. For example, the Fiscal Strategy Report (2011) assumes that top-down constraints on government expenditure (operating allowances) continue into the medium term.