Uncertainty
As noted above, there is much uncertainty about the indicator estimates. There are two broad sources of uncertainty which can lead to revisions in the indicator estimates:
- estimation uncertainty of the key model parameters (ie, the output gap and the average sensitivity of tax revenues to changes in the output gap), and
- forecast uncertainty relating to future fiscal and economic developments.
In addition, summary indicators such as fiscal impulse do not take account of the composition of fiscal policy changes or how a change in fiscal policy will be transmitted through the economy. Treasury research using time series statistical analysis indicates that spending and taxes have different effects on New Zealand GDP.[3] Therefore the fiscal impulse indicator is only a very imprecise guide to the impact of fiscal policy on the economy.
Sensitivity analysis is performed by calculating the indicators using alternative output gaps (from the RBNZ, IMF and OECD) and values for the elasticity of tax revenues with respect to the output gap which are half and twice the magnitude of the baseline estimate. The range of alternative estimates is plotted in Figures 4 to 6 (with data reported in Tables 16 and 17). Differences in the output gap estimates are mainly the result of differences in estimation technique, although it also reflects different institutions' judgements about the forecast outlook and the availability of data at the time of forecast finalisation. Accordingly, it provides an indication of uncertainty due to model specification but it does not capture total forecast uncertainty.
An alternative means of illustrating uncertainty is to show a probability distribution around the central forecast. A probability distribution requires making some assumptions about future forecast errors based on historical forecast errors of observable economic and fiscal variables and historical revisions to the Treasury's output gap estimates. In Figure 3, a fan chart of the cyclically-adjusted balance indicator is shown. The probability intervals calculated are conditional on current policy and reflect historical revisions to the Treasury's official output gap estimate, rather than the full uncertainty implied by different estimation techniques. Details of the methodology and parameter values for the confidence intervals are reported in Treasury Working Paper 10/08.[4] This analysis would suggest that there is a structural fiscal deficit in 2011/12 with a high degree of confidence. There is considerable uncertainty in the forecasts over a three-year horizon.
- Figure 3 - Fan chart for cyclically-adjusted balance

- Source: The Treasury
Note: the bands represent sequential deciles such that the difference between the 10th and 90th percentiles represents an 80% confidence interval.
- Figure 4 - Output gap range

- Source: The Treasury
- Figure 5 - Cyclically-adjusted balance range

- Source: The Treasury
- Figure 6 - Core Crown fiscal impulse range

- Source: The Treasury
Notes
- [3]Iris Claus, Aaron Gill, Boram Lee and Nathan McLellan (2006) "An empirical investigation of fiscal policy in New Zealand." New Zealand Treasury Working Paper 06/08 http://www.treasury.govt.nz/publications/researchpolicy/ wp/2006/06-08/. The degree to which the fiscal impulse indicator matches the time series estimates depends on the exact form of the latter. In neither of the time series specifications does the summary indicator match the time series estimate across the entire sample period.
- [4]Oscar Parkyn (2010) “Estimating New Zealand's Structural Budget Balance.” New Zealand Treasury Working Paper 10/08 http://www.treasury.govt.nz/publications/research-policy/wp/2010/10-08/.

