4.5 Fiscal stabilisation outcomes
New Zealand’s recent experience with fiscal stabilisation can be divided into an assessment of the automatic stabilisers and of discretionary fiscal policy changes.
The degree to which taxes and transfers co-vary with the level of economic activity is critical to determining the degree of automatic fiscal stabilisation. The degree of progression in New Zealand’s income tax structure means that tax revenue increases by a greater proportion than the increase in GDP (Fowlie, 1999; Tam and Kirkham, 2001). Over the period 1987-2005 New Zealand’s tax and expenditure system resulted in “reasonably pronounced counter-cyclical behaviour of the government primary (non-interest) deficit” (Buiter, 2006, Figure 23, page 50). But the presence of automatic stabilisers does not guarantee counter-cyclical policy.
Assessing the extent to which discretionary fiscal policy has been pro-cyclical or counter-cyclical poses significant challenges. Information on whether changes in the economy are cyclical or structural will always be limited. Recent developments in modelling techniques endeavour to allow for the effects of changes in the size of fiscal initiatives, the composition of expenditure and taxation (for example, Blanchard and Perotti, 2002), the level of public debt (for example, Chung and Leeper, 2007; Favero and Giavazzi, 2007) and initial macroeconomic and fiscal conditions (for example, Perotti, 1999; Romer and Romer, 2007a).
The Government’s operating surplus has risen markedly since the 1990s. This balance is not sufficient to assess the impact of fiscal policy on domestic demand. The Treasury tends to use a suite of fiscal indicators to assess the likely impact of current fiscal policy on the business cycle. One indicator is a fiscal impulse measure based on the change in a cyclically-adjusted fiscal balance (Philip and Janssen, 2002). This indicator attempts to gauge the initial (first round) impact of discretionary fiscal policy (measured by the change in the structural balance) on aggregate domestic demand. Data for this indicator is drawn from the Statement of Cash Flows and the fiscal balance used to derive the indicator is the difference between primary cash from government operations and primary cash payments to operations plus capital cash spending.
Based on current economic information, this measure implies that fiscal policy has tended to restrain short-run demand in most years since 2000, with small injections to demand in 2004 and 2006. However, as the size of the structural surplus is forecast to decline over coming years, this indicator implies that fiscal policy is expected to become less of a constraining influence, with a reasonably significant fiscal impulse indicated for 2007 (see Figure 5). Forecasts of the fiscal impulse have, however, been subject to significant reassessment over recent years.
As Philip and Janssen (2002) recognise, this approach to assessing the impact on domestic demand needs to be viewed with caution. It is designed simply to assess the initial fiscal injection (or contraction) to demand. It does not take account of second round effects, changes in the composition of the fiscal balance, and the way private expectations can influence responses to a fiscal pulse. Furthermore, recent work by Romer and Romer (2007a), using a narrative approach to identify discretionary structural changes to fiscal policy in the USA, suggests that the traditional cyclical adjustments to budget data have a tendency to overstate the derived structural changes in the budget. If this is correct, the fiscal impulse measure may be overstating the size of structural budget balance changes and therefore the size of the derived contraction (or impulse) to domestic demand.
An alternative approach to assessing the impact of fiscal policy on New Zealand GDP using vector-autoregressive (VAR) modelling has been developed by Claus, Gill, Lee and McLellan (2006). This approach attempts to measure both the initial (first round) effects and the induced effects of discretionary fiscal policy on domestic demand by explicitly estimating and incorporating private sector responses and response lags. It also takes account of compositional effects by estimating separately the impact of changes in government expenditure and government revenue on GDP. Data for this VAR-based approach is drawn from Statistics New Zealand’s quarterly National Accounts (using SNA93 standards). Although the data are derived on a different conceptual basis from the cash-based series used by Philip and Janssen, the fiscal balance measures are similar.
Refinements of this vector-autoregressive approach are currently being investigated in the Treasury. The results from this VAR work to date suggest that fiscal policy has generally been pro-cyclical during the late 1990s and from 2001 to 2003, and has tended to be pro-cyclical again in 2004 and 2006 (see Figure 6). The estimated size of the fiscal impulses is often smaller and the pattern quite different from that implied by the traditional Treasury measure of fiscal impulse (see Figure 7). This difference reflects, in part, differences in the way discretionary policy changes are identified and the fact that the VAR approach applies different weights for the expenditure and revenue impacts on domestic GDP. When the expenditure and revenue components of the traditional fiscal impulse indicator are weighted by the multipliers derived from the structural VAR model, the size of the traditional fiscal impulses tend to be smaller, although the direction of changes still differ from the fiscal VAR impulses in some periods.
Differences also arise from the fact that the VAR approach explicitly captures the interactions between fiscal variables and the business cycle and explicitly captures some, but not all, fiscal composition effects. Composition effects are likely to have been important. For instance, much of the recent growth in government expenditure has been in areas where a significant demand impact can be expected, such as consumption of non-tradable goods and services, wages of public sector employees and transfers to low and middle-income households (for example, the Working for Families package).
Furthermore, there are various reasons to think that an assessment of the impact of growing taxation revenue growth should take account of the circumstances explaining the taxation revenue growth and how individuals and firms may respond to taxation revenue growth. If households tend to be more likely to be finance-constrained than firms, then it would be important to distinguish between the sources of taxation revenue growth. The strong rise in the government operating surplus in recent years has been in part the consequence of an unexpectedly strong growth in company tax revenues. This taxation growth may have less of a constraining influence on business investment and domestic demand than, say, an equivalent growth in personal income taxation revenue would have on household consumption demand if households are more likely to be finance-constrained than firms.
Similarly, as Dunstan, Hargreaves and Karagedikli (2007) argue, this increase in corporate tax revenue may not have had as large an effect on domestic demand as an equivalent increase in government expenditure. Indicators suggest that government expenditure growth has been strong over recent years. Figure 8 shows government investment expenditure has gradually increased over recent years. Table 2 reveals the relatively strong growth in public sector employment in recent years. This has occurred during a period of very low unemployment and strong private sector demand for labour.
Hence, although the practice of Government setting and publishing short-term intentions is aimed at supporting macroeconomic stability and although Government has recently shown an increasing willingness to adjust new expenditure plans in response to macroeconomic conditions, there is contrasting evidence concerning the actual outcome of discretionary fiscal policy on the business cycle. Improving understanding of the impact of fiscal policy on domestic demand and the business cycle is an area worth further investigation and it is an area in which Treasury is devoting more research. The recent paper by Dungey and Fry (2007) is part of that process.
