GST compensation trades one inefficiency for another and won't achieve 'fairness'

The GST debate is on again in earnest. Prime Minister Malcolm Turnbull deserves kudos for bravery, if nothing else.

The GST is a regressive tax. Any proposal to increase the rate of GST or expand its application will impact lower income earners disproportionately. Since the 2014 Budget, public perception is that the government has treated low income earners unfairly. Proposing to increase a regressive tax is, therefore, a bold strategy.

Many suggest the GST’s regressive effect could be mitigated by compensating low income earners. But how much compensation should be paid, and to whom?

Poorly defined proposals

Assessing proposals for compensation is difficult. At present they are poorly defined. The most developed compensation proposal is NSW Premier Mike Baird’s. It would see households which earn under $100k “fully compensated”. This invites four questions. Firstly, why $100,000? Secondly, is that income before tax? Thirdly what counts as a household? Finally, how much will this cost?

$100,000 is a nice, round figure; but what does it represent? In gross - that is pre-tax - income $100,000 per year represents the top of the 60th percentile of income distribution. The rationale for extending and increasing the GST is to expand the revenue base.

Baird’s proposal would collect increased revenue from 60% of households only to transfer it back to them as compensation. The technical term for this is tax churn – it makes compensation for GST increases a modern-day task of Sisyphus. Tax churn is the cost of collecting the tax, then handing it back. It is an inefficiency to be avoided.

If we wished to restrict tax churn and only pay compensation to the lowest 20% of household incomes, we would have to cap compensation at about $35,600. This figure is only slightly higher than the national minimum wage. Such a proposal would test the limits of even Turnbull’s bravery.

Considering pre-tax income raises another complication. In Australia incomes are taxed individually. Assessing GST compensation by “household” leads to strange outcomes. For example, a couple household with a single pre-tax income of $100,000 will usually have a post-tax disposable income of about $75,000. Yet if both partners work, each earning $50,000, each pays a lower rate of income tax. Such a household would have post-tax disposable income of around $84,400. On its face, Baird’s proposal would compensate these households equally. This seems unfair.

Alternatively, householders could be reimbursed for GST paid. Accounting for the actual GST paid would reduce the disparity compared to assessing compensation on pre-tax income. However, such a system would be cumbersome and expensive to administer – taxpayers would need to keep records of all purchases which attracted GST. Most people – like Mitch Hedberg – don’t tend to keep receipts for donuts.

Making households equivalent

What counts as a “household”? A couple without children sharing an income of $100,000 are different to a couple whose same income supports four school-aged children. Is an individual living alone a “household”? If so, any GST compensation rules may be a strong incentive to live apart.

Presumably Baird means “equivalised household income” - the measure used by the Australian Bureau of Statistics that normalises income for a given household size, to allow fair comparison. This would mean giving households different compensation based on a set of criteria – size, work status, dependants, for instance – as the ABS “equivalised household income” data does.

It would invite questions about what other criteria might be relevant; perhaps geography, education level or net wealth. Such an equalising exercise – and its administration – is not impossible. However, it is complicated and therefore expensive. It will also be, necessarily, an imperfect measure.

The cost

Finally, how much will this compensation cost? As the proposal is imprecise, answering this question is difficult. But consider this; Parliamentary Budget Office (PBO) modelling suggests that a 15% GST expanded to extra goods and services would raise an additional $65 Billion before compensation in 2017-18. Premier Baird’s proposal would, 15 years from now, raise $35 billion annually, after compensation. This comparison is imperfect. However, these figures demonstrate the huge quantum of compensation Baird proposes. The gap between what the PBO suggests could be raised next year and what Baird proposes to raise in 2030 is larger than federal spending on education; about $30 billion in 2014-15.

The GST is a blunt instrument. It does not discriminate. This is one of its strengths; it is a difficult tax to avoid. Yet this poses a dilemma for reformers. The complicated rules on zero-ratings for food, health and education represent an imperfect attempt to make the GST fairer. Proposing to reform the GST and retain this “fairness” is fraught. Such proposals merely substitute an inefficient and complicated method for collecting the tax with an inefficient and complicated method for redistributing it.

This difference is not academic. Inefficient collection of the GST risks under-taxing high income earners. Inefficient redistribution risks under-compensating low income earners. If our concern is fairness, the former seems preferable.

Some suggest that GST reform should fund cuts to income and company tax. This is not Baird’s proposal. Given his proposed compensation package, the GST increase could not fund both health spending and additional tax breaks. However, such cuts appear to be Treasurer Scott Morrison’s preference.

Income tax cuts would only be affordable if we reduced the number of people fully compensated for the GST increase. This would mean increasing the tax burden on low income earners so that we could decrease the tax burden on higher income earners. If we are concerned with taxing people according to their capacity to pay – what tax academics call vertical equity – these proposals seem deeply unfair.

Yet, increasing the GST can be fair. High GST does not predict national income inequality. In 2012 Australia had a Gini co-efficient – which measures income inequality – of .326 after taxes and transfers. (A Gini co-efficient of 0 represents perfect equality and 1 represents perfect inequality). Denmark’s was .249, Sweden’s .274. Those countries assess GST at a rate of 25%, though Sweden assesses food at 12%.

Sweden and Denmark collect more tax overall than Australia. In 2010 Sweden’s tax take was about 45% of GDP, Denmark’s about 48%. Australia’s was 25.6%. The evidence from Sweden and Denmark is that high consumption taxes and low income inequality are compatible. However, this requires collecting more tax and redistributing it progressively through social programs targeted at those in need, not merely as “compensation”. Such programs require administration and therefore cost. But it is a preferable cost to mere tax churn.

If Turnbull wishes to demonstrate true political bravery, he might propose that strategy.

Tomas Fitzgerald has received funding from the WA Bar Association. He is a member of the NTEU and WA Labor.