Statement 5 (continued)
The Government's receipts estimates are generally prepared using a 'base plus growth' methodology. The last known outcome (2012‑13 for the 2014‑15 Budget) is used as the base to which estimated growth rates are applied, resulting in receipts estimates for the current and future years.
Most of the indirect heads of revenue, such as GST and fuel excise, are forecast by mapping the growth rate of an appropriate economic parameter directly to the tax growth rate in the relevant head of revenue. A number of income taxes also involve determining whether this tax will be paid in the year the income is earned, such as for pay‑as‑you‑go withholding tax, or in future years, such as for individuals' refunds.
The Government's receipts forecasts, like all forecasts, are subject to a margin of error. The receipts forecasts between 2002‑03 and 2007‑08 under‑predicted outcomes (Chart A1). For example, the 2007‑08 Budget forecast taxation receipts to grow by 5.0 per cent in 2007‑08, while the outcome was growth of 8.1 per cent, a forecast error of 3.1 percentage points. Since 2008‑09, the outcome for receipts has been lower than the Budget forecast, broadly reflecting the effects of the GFC and subsequent weaker than expected growth in wages and prices.
These historical forecasting errors form the basis of the confidence intervals for the revenue forecasts presented in Appendix B of Statement 3.
Forecasting errors may be split into three underlying sources: errors in economic forecasts that underpin the receipts forecasts; errors in translating the economy to receipts forecasts; and miscellaneous factors such as post‑Budget policy decisions, court decisions regarding tax law interpretation, changes in compliance activities of the ATO, and revisions to historical economic data. Note that there may also be secondary errors relating to the timing of payments of tax — even if the underlying forecasts were accurate, revenue may be recorded in the fiscal year before or after it was expected.
The relative contribution of each of these sources of error varies from year to year, but there is generally a strong correlation between the accuracy of the forecasts of the nominal economy and the forecasts for tax receipts, as shown in Chart A2.
Chart A2 plots the forecast errors for nominal GDP against the errors for tax receipts. It shows where there has been an underestimate of nominal non‑farm GDP growth, tax receipts tend to be underestimated and vice versa.
Furthermore, on average, economic forecasting errors will be magnified in receipts forecasting errors, due to the progressive nature of personal income tax. The lower and upper lines are based on aggregate elasticities (of receipts with respect to nominal non‑farm GDP) of 1.0 and 1.5 respectively, assuming an error of plus or minus 0.5 per cent if there is zero error on the economic forecasts. Forecasting errors outside this range could be a result of factors such as timing of tax receipts.
For example, in 2002‑03, nominal non‑farm GDP growth turned out to be around 1 percentage point higher than forecast, but growth in tax receipts (excluding CGT) was almost 4 percentage points higher than forecast. That is, the error in the revenue forecast was higher than the around 2 percentage points that the rule of thumb suggests should be theoretically associated with an economic forecasting error of that magnitude.
The forecast for 2013‑14 tax receipts (excluding CGT) in the 2013‑14 Budget is expected to be an over estimate of around 3.9 percentage points, compared to an over estimate of around 1.1 percentage points for nominal non‑farm GDP growth. The forecasting error is partly attributable to the shortfall in resource rent taxes. The resource rent tax base is not expected to relate as closely to nominal GDP as taxes on wages or corporate profits. The largest driver of the expected forecast error for 2013‑14, however, is gross income tax withholding (mostly withholding taxes on wages), which is estimated to be nearly $7 billion (4.1 per cent) short of the forecast in the 2013‑14 Budget. This result, caused by much lower than expected growth in average wages, comes after five years of very accurate forecasts for gross income tax withholding (average error of 0.6 per cent).
Discussions of earlier years can be found in previous budgets.
From 2008‑09, forecasting errors in tax receipts have been significantly affected by the economic downturn related to the GFC, particularly with regards to capital gains tax (Chart A3).
Forecasting CGT is very difficult for several reasons. First, price movements above or below the assumption may cause CGT to be significantly different from the forecast. Secondly, CGT only applies to realised gains, so even if the asset prices are consistent with the assumptions, there may be more or less gains realised than was assumed. Thirdly, following the GFC, a large stock of capital losses were carried forward (see Box 2 of Statement 5 of the 2010‑11 Budget), and the utilisation of these losses generates large uncertainty in both the timing and magnitude of the forecasts. Finally, relevant data on CGT is available over a year after the event, meaning that the forecasting assumptions will be slow to respond to changing taxpayer behaviour.
In light of concerns regarding the performance of economic and revenue forecasts, the Secretary to the Treasury commissioned an independently overseen Review of Treasury Macroeconomic and Revenue Forecasting, which made four recommendations in regard to revenue forecasting (out of 11 recommendations in total). These four revenue‑specific recommendations were:
- developing a three sector company tax model (mining, financial and other);
- investigating whether further information can be drawn from the Australian Taxation Office's liaison with large corporate taxpayers;
- constructing a micro‑simulation model for forecasting personal income tax; and
- giving further consideration to the appropriate balance between the top‑down versus bottom‑up approaches to forecasting revenue.
These have been implemented or are in the process of being implemented.