Intergenerational Report misses opportunity to lay groundwork for Tax White Paper

The Intergenerational Report (IGR) is a valuable contribution to public policy, but there is scope for it to be significantly improved. While it has been subject to some criticism, the real issue missing from the IGR is modelling of one entire side of the Budget – the tax side. As a result, an opportunity has been missed to enable a more informed discussion for the Tax White Paper.

The IGR basically assumes that total tax revenue remains at the same proportion of the economy (23.9 per cent) for the whole period 2020‑21 to 2054-55 (see IGR, p113). This is similar to the approach taken in all IGRs since the first one in 2002.

As it is, the IGR is only half useful. It usefully models spending, but it largely misses tax.

This causes a plenty of problems.

The IGR provides no useful information on what is happening with taxes over the long term. We don’t have forecasts of individual taxes, or forecasts of tax expenditures. We don’t know who will bear tax burdens over coming decades. Will it be the young or the old? Families or singles? Rural or urban Australians? Small or large businesses? We don’t know how much each industry will pay in tax; we don’t have information on the tax on particular goods or services of interest (eg internet trade); and the IGR can’t separately estimate the impact of ageing or migration on tax revenue. The IGR has a general discussion on tax and demographics (see section 2.3) but does not conduct a substantial analysis.

It should be noted that the IGR doesn’t do projections for spending by demographic group, let alone tax. It would be improved by doing demographic analysis for spending alone; but the largest improvement would be from doing this analysis for both spending and tax.

The IGR also doesn’t have information on whether some taxes are forecast to decline, for example due to globalisation: all weaknesses in one tax are assumed to be replaced by increased revenue from other taxes. Conversely, the IGR assumes a revenue boom, for example due to bracket creep or another mining boom, is fully offset by an equal reduction in personal or other taxes.

Modelling on all of these issues would all be very useful for the debate around the tax reform White Paper.

Another concern is that the treatment of tax in the IGR is inconsistent with the modelling of spending. Unlike tax, the IGR separately models each major item of government spending. If spending on medicines or hospitals increases, the IGR doesn’t assume that welfare spending or education will fall to fully offset this. Of course, if there were a full offset then the IGR would be a pretty unhelpful document: both spending and revenue would remain fixed forever at a specified proportion of the economy.

The IGR also mis-states the future budget situation. If taxes were individually modelled, this may show revenue growing faster or slower than assumed in the IGR, and future deficits may be larger or smaller. If improved modelling shows revenues are set to decline, then the budget problem becomes both larger and more urgent. And a decision to maintain revenue as a portion of GDP will need to be done much more explicitly, rather than assumed as in the IGR: a debate needs to be had about which taxes will need to be increased or broadened.  Conversely, if revenues are set to increase as a proportion of GDP, then the Budget emergency is less important. Either way, Australians should know.

Why doesn’t the IGR model each tax individually, similar to the approach for spending? It should be no harder to model taxes than spending on medicines, education or welfare, all of which is done in the IGR.

The reason why the IGR takes this approach is that modelling of personal tax will show revenue growing very quickly if there is no indexation of tax thresholds: revenue would probably grow faster than GDP (see IGR, p115). This would be pretty unrealistic (and inefficient) future, as the then Secretary to the Treasury, Dr Martin Parkinson, highlighted in several speeches last year. But the IGR’s response to this problem is fairly inelegant, causing all the problems raised earlier.