Imagine if everyone with a surname starting with the letter C didn’t have to pay income tax. For some arcane reason, back in the mists of time when the tax was introduced in Australia, those with a ‘C’ name were completely exempted, and the exemption remained on the books, stubbornly resistant to efforts to remove it.

This exemption for people with ‘C’ names would be classed as a ‘tax expenditure’. This is what it’s called when some group of people, or some class of activity, is taxed at a lower rate than others in similar circumstances (or is not taxed at all). They’re called tax expenditures because in the end there’s not a huge difference between exempting everyone with a ‘C’ name from income tax and just sending them all a cheque. That’s why tax concessions are sometimes described as “hidden spending” or “spending through the tax code”.

There’s often a good reason to provide a tax concession. For example, war veterans’ pensions generally aren’t taxed, and nor are “particular World War II-related payments for persecution”. Fair enough, I say. But people receive a benefit from tax concessions, and providing them reduces government revenue, so it’s a good idea to keep them on the public record and detail how much they’re worth. This enables people to debate whether particular concessions are a good policy idea, or whether the revenue that’s foregone in providing them could be better put to some other use. As George W. Bush’s chief economic adviser put it, “it is just as important to focus on stealth spending implemented through the tax code as on explicit spending.”

But not everyone sees tax concessions as “stealth spending”. To some, the money that isn’t taxed belongs to the individual, not to the government, so it isn’t right to view the absence of a tax as a concession. The close examination of tax expenditures, Judith Sloan suggests, can lead down a slippery slope in which it’s thought that “100 per cent of all income should belong to the government and any residual that taxpayers are allowed to keep should be regarded as a concession”.

I think this is hyperbole. Suppose that a federal government proposed to close my imaginary income tax loophole and subject everyone to the same rules regardless of the first letter of their last name. This act of alphabet warfare would raise revenue that could be used to cut taxes for everyone, or to boost spending on all those worthy things that governments spend money on. Ending the exemption would also make things fairer: after all, why should all the Browns and DiPierdomenicos of Australia pay tax while the Cowgills get off scott free?

Do you think it would be right to view the alphabet-based exemption as an unjustified tax expenditure, eroding the revenue base and undermining equity for no good reason? Or do you think that applying income tax to the Cowgills should be seen as the illegitimate encroachment of a bloated state onto its citizens’ private property? I tend towards the former view, as you might have guessed.

All of this might seem a little arcane, but it’s really at the heart of the superannuation debate Australia has been embroiled in for the past week or two. That’s because the concessions associated with super are, collectively, the biggest tax expenditures in the Australian system. This year, the Treasury estimates that super concessions will be worth around $32 billion to the people who receive them, and this will rise to around $45 billion by the middle of the decade.

That’s a lot of money. About 20% of the benefit is enjoyed by people in the top 5% of the income distribution, while the top 1% receive around 5.3% of the benefit. I think there’s grounds for reducing the tax expenditures on super both in order to increase revenue and to reduce the unfairness of the system.

Tax expenditures on super are large, and they flow disproportionately to people at the top end of the income scale. But it’s important not to fall into the trap of thinking that Government could raise $32 billion by ending these concessions, or the trap of thinking that they should completely eliminate them.

The first mistake represents a misunderstanding of what the tax expenditure figures mean. The $32 billion figure is an estimate of how much the super concessions are worth to the people who receive them. If government ended the concessions tomorrow, it wouldn’t raise $32 billion. Instead, people’s behaviour would change – they’d put less into super, or take more out, and avoid paying the tax that way. Treasury thinks that the two main super concessions are worth around $30 billion to the people who receive them in 2012-13, but it estimates that tax revenue would only rise by around $23 billion if the concessions were removed. The difference between these two figures is due to the change in behaviour that would happen after the rules are changed.

The second error, at least in my view, is believing that tax expenditures on super should be completely removed. Super concessions lie somewhere between the income tax exemption for ‘C’ named people and the exemption of payments to people persecuted during WWII. They’re there for legitimate reasons – to promote saving for retirement and as compensation for the compulsory nature of the scheme – but the size and distribution of them is difficult to justify. I think that super concessions should be scaled back a little, particularly for high income earners, to improve the equity and sustainability of the system, but that doesn’t mean that they should be cut to nil.

The tax expenditure figures can be misused on both sides. The existence of a concession doesn’t mean, in and of itself, that there’s a loophole that should be closed, and the amount of money that would be raised by doing so is often overestimated. On the other hand, a concession is a concession, not a “concession”. Just like the exemption of people with ‘C’ names from income tax, treating some people or types of activity differently from others amounts to much the same as just giving them a cheque, so tax expenditures should be closely scrutinised and justified.