What if GST and payroll tax are unconstitutional?

In my second-round submission to the Henry tax review, I argued that the collection mechanisms for GST and personal income tax violate s.82 of the Constitution, which says in part: "The costs, charges, and expenses incident to the collection, management, and receipt of the Consolidated Revenue Fund shall form the first charge thereon ..."

At face value, those words would make it unconstitutional to require businesses, at their own expense, to collect income tax payable by employees or GST payable by customers. Tax-deductibility of the expense is insufficient because it still leaves an after-tax component that is not charged to Consolidated Revenue.

In the case of the GST, sellers explicitly collect the tax from buyers by issuing tax invoices. So the obvious way to make the GST comply with s.82 is to eliminate tax invoices and make the tax payable on sales instead of purchases. But that raises another problem: how do we handle input credits, for which tax invoices serve as proof of payment?

One solution is to remove restrictions on credit for GST paid on business inputs. In general, whether GST has been paid on inputs depends on the nature of the inputs. But whatever their nature, the GST is presently reclaimable only for inputs purchased from GST-registered suppliers, i.e. suppliers who presently issue tax invoices. The GST paid by "input-taxed" suppliers and passed on in their prices isn't reclaimable even by GST-registered customers; it's called "sticky" GST, and leads to tax on tax. If one could claim input credits for both classes of suppliers, one wouldn't need tax invoices to distinguish between them.

This solution would slightly reduce the GST base, but would substantially cut compliance costs for traders whose turnover is low enough to qualify for input-taxed status, but who are forced to register for GST just because prospective customers want to claim input credits. (The number of affected micro-enterprises is reportedly in the hundreds of thousands.)

A better solution, yielding a far greater reduction in compliance costs with no contraction of the GST base, is to do away with input credits by making the GST a retail tax instead of a VAT.

Conventional "wisdom" holds that a VAT is harder to evade, because sellers who fail to declare sales may be exposed through their customers' input credits. That's complete nonsense because any purchase that qualifies for an input credit will also qualify for an income tax deduction, which is more valuable and therefore more likely to make customers keep records of their purchases.

And what if you're a retailer, so that your usual customers don't claim input credits? In South Korea they solve that problem by offering tax incentives to retail customers who use credit/debit cards, which create electronic records of purchases. This would work just as well for a retail tax as for a VAT. If the "incentives" were rebates rather than deductions, they would work even in the absence of an income tax.

Conventional "wisdom" further argues that a VAT encourages traders to declare sales so that they can claim the corresponding input credits - conveniently forgetting that income tax has the same feature, and that it is more lucrative to hide both sales and purchases than to declare both. The only merit in this argument is that traders will pay VAT on any undeclared inputs - provided of course that the non-declaration has not been suborned by suppliers who also evade VAT!

But the clincher is that a VAT opens a second front for evasion: bogus input credits. Under a VAT, traders who overstate their inputs get income tax deductions plus VAT credits, whereas under a retail tax they only get income tax deductions.

In truth, the one thing a VAT does better than a retail tax is to generate compliance costs, which give larger businesses an advantage over smaller ones and deter start-ups that might compete with established players. As politicians are beholden to big business for campaign funds, the best hope for small business may be a constitutional challenge to the GST as presently implemented.

Either of the above remedies - relaxing restrictions on input credits or switching to a retail tax - would be within the Henry review's terms of reference, which rule out raising the GST rate or broadening the base but say nothing about narrowing the base or changing the implementation.

Raising the GST rate in order to replace that other broad-based indirect tax, namely payroll tax, would exceed Henry's terms of reference and breach the letter of the Rudd Government's election commitment. But it would more than satisfy the spirit of that commitment, because consolidating indirect taxes would reduce compliance costs, hence the cost of living (which is inflated not only by indirect taxes but also by their compliance costs). Besides, the Government's hand might be forced by s.90 of the Constitution, which forbids the States to impose excise taxes.

In the last relevant High Court case, namely Ha v. NSW (1997), the majority held that an excise is "an inland tax on a step in production, manufacture, sale or distribution of goods". If paying the workers is such a "step", that definition would sink payroll tax.

(The same definition, by the way, would also scuttle the existing State stamp duties on new cars and sales of livestock. The States like to live dangerously.)

The three dissenting judges preferred a narrower definition of excises. In their view the purpose of s.90 was to "prevent impairment by the States of the common external tariff," so that "A State tax which fell selectively upon goods manufactured or produced in that State would be an excise duty ..." They added: "Whether a tax which falls upon locally produced goods discriminates against those goods in favour of imported goods is a question of substance, not form." In substance, domestic payroll tax on labour embodied in goods manifestly discriminates against locally produced goods because it is not levied on the corresponding labour embodied in equivalent imported goods.

Does it matter that payroll tax affects services as well as goods?

Not under the majority definition, because the GST also affects services, and nobody is suggesting that the States could have imposed the GST; that's why the Commonwealth imposed it and handed over the revenue. Under the minority definition, the question is whether a law forbidding discrimination against local goods can be circumvented by discriminating against local services as well. One should hope not. (But I'm not a lawyer and this article is not advice!)

Admirers of payroll tax are keen to point out that it is "shifted" downstream in higher prices, like the GST. They are not so keen to admit that both taxes are also shifted upstream, where GST affects production in general whereas payroll tax affects employment in particular.

In an appendix to the Treasury paper "Architecture of Australia's Tax and Transfer System", it is shown mathematically that an all-in payroll tax is equivalent to an all-in consumption tax under the following unrealistic conditions: (i) no initial savings or capital; (ii) no returns on investment above the discounting rate; and (iii) no exports or imports.

The corresponding realities are: (i) pre-existing assets add to the consumption base; (ii) so do super-normal returns on investment; and (iii) a consumption tax exempts exports and hence the local labour embodied therein, whereas a payroll tax exempts foreign labour embodied in imports (although the Treasury paper is a bit coy on this point).

Consequently a payroll tax has a narrower base and, lo and behold, does more damage to local employment.

But what if payroll tax really were a consumption tax? A GST also taxes consumption. So if a State GST would be unconstitutional, where does that leave payroll tax?

If State payroll taxes were struck down, the Commonwealth could of course impose its own payroll tax and distribute the revenue to the States. But that would be political madness, because it would waste the best-ever opportunity to eradicate the hated payroll taxes (think of the kudos!), and because it would worsen vertical fiscal imbalance, so that the States could more easily blame Canberra for underfunded services.

Let me therefore offer an alternative solution that gets rid of payroll tax and reduces or eliminates vertical fiscal imbalance. One advantage of a retail tax is that the absence of input credits makes it easy to have different rates (on a uniform base) in different States. (In this context the "States" will be taken to include the Territories.) So let the retail tax rate in each State be set annually by the Federal Parliament at the "request and consent" of the State Parliament, and let the revenue so raised in each State be refunded to the State on the condition that the State refrains from imposing certain other taxes (e.g. income tax).

Does this arrangement amount to a State excise tax in violation of s.90 of the Constitution? No, because it's imposed under Federal legislation. Does the variation in the rate amount to discrimination between the States in violation of s.51(ii) and s.99? No. The Commonwealth invites each State to pass a "request and consent" act. The States respond as they see fit and the Commonwealth meekly accepts their responses. Where's the discrimination?

In the unlikely event that "discrimination" is an issue, there's another way to legitimise the same arrangement. According to the "minority" definition of an excise, the States can impose their own retail taxes, in which case they can refer the collection power to the Commonwealth under s.51(xxxvii), and the Federal Parliament can authorise the collection by the ATO on the condition that the referring States accept a uniform base and refrain from imposing certain other taxes.

The enabling Federal and State legislation could cover both constitutional options, so that different High Court judges could approve the arrangement for different reasons. As long as a majority of judges agree that the arrangement is constitutional, it doesn't matter if they disagree on the reasons!

In proposing a retail tax as a constitutional fix, I have not been wearing my Prosper Australia hat. But now let me put it on to answer just one question: if we must have a retail tax, and if the need to replace payroll tax provides some wriggle-room concerning the rate and the base, how should the retail tax treat real estate? My answer has four parts.

1. To loosen up the supply chain, all property sales (some of which are presently subject to GST) should be exempt.
2. To encourage construction, buildings should be exempt, leaving only sites subject to tax.
3. To maximise the incentive to build commercial accommodation and seek tenants, commercial landlords (who presently collect GST on rents received for sites and buildings) should pay retail tax on the imputed rental values of their sites alone, regardless of whether the sites are developed or let to tenants; and any contractual provisions requiring tenants to pay the tax (or increments in the tax) should be void.
4. To maximise the incentive to build rental housing and seek tenants, residential landlords (who are presently input-taxed) should be treated as in (3).

A retail tax implemented in this way would improve housing affordability and provide a stimulus for construction and related industries. And regardless of how it treated real estate, it would reduce vertical fiscal imbalance, eliminate payroll tax, improve international competitiveness, create jobs, and reduce compliance costs that feed into the cost of living. Anyone who can bring this about through a constitutional attack on the existing system will do the country a great service.