Today’s post by Lisa Martin plots a summary of historical National Treasury revenue forecast errors by component. As we showed in our paper on fiscal forecast errors, between fiscal years 2016/17 and 2020/21, these surprises reflected generally weaker than-expected personal income tax (PIT), corporate income tax (CIT), and value added tax (VAT) receipts. Post-pandemic, PIT, CIT, and domestic VAT tended to be under-estimated, initially reflecting higher commodity revenues from unexpectedly high global commodity prices and an increase in mining exports before becoming more broad-based. To summarise the overall impact of government revenue forecast errors, today’s post plots the cumulative totals of forecast errors by category since 2013. The largest
individual over-estimates of revenue over this period were for PIT and VAT, particularly in the 2020/21 fiscal year. The largest under-estimations of revenue components occurred in the 2021/22 and 2022/23 fiscal years, with CIT being the largest contributor to underestimations of revenue. With regards to the cumulative effect of forecasting errors, PIT and VAT underperformed relative to forecasts, while CIT exceeded expectations over the period. Large PIT over-collection reflected a short-lived increase in commodity prices and mining exports in 2021 and 2022.
Footnote
In the chart above, the T-1 label shows forecast errors for in-year ‘revised estimates’. The labels T, T+1, T+2 compare outcomes from the latest Budget Review (the 2026 edition) to the three Medium Term Expenditure Forecast years published in earlier Budget Reviews. For example, the
outcome published for 2024/25 in the 2026 BR is compared with
previous estimates/forecasts to create the ‘T+2’ (grey) bar at 2022/23, ‘T+1’ (red) bar at 2023/24, ‘T’ (light blue) bar at 2024/25 and the ‘T-1’ (dark blue) bar at 2025/26.
EconData makes it easy to assess fiscal forecast errors, contact us for a demo.
