While Brendan Coates makes some valid criticisms of Paul Keating’s recent claims about Australia’s retirement income system, his own claims and assumptions (and those of the Retirement Income Review) are also open to question.
First, Coates is right that Keating exaggerates the cost of the age pension: Australia’s age pension costs a fraction of the amount most OECD countries face today, and those costs are forecast to fall as a percentage of GDP whereas others’ costs are forecast to increase.
But Coates’ and the RIR’s claim about the cost of superannuation tax concessions is equally exaggerated. They define the concessions against a benchmark where both contributions and earnings would be taxed at the employee’s marginal income tax rate. Such a tax regime would impose an enormous penalty on savings, particularly long-term savings, undermining people’s choice as to when they consume their income; it would also be totally inconsistent with the retirement income system’s basic design which compels people to defer consumption in order to self-fund their retirement.
It is wrong to present such ‘tax concessions’ as a major cost akin to that of the age pension. There is a legitimate debate about the precise level of tax that should apply to superannuation savings, but the post-Turnbull tax regime is probably about right if rather messy.
Secondly, Coates is also right that Keating should not be claiming that increasing the SG would have no effect on wages. At best Keating’s claim might have some validity in the short-term while wages growth is currently low, though only at the risk of increased unemployment.
But Coates should equally accept that increasing the SG would have a benefit not just a cost, the benefit being increased retirement income. The challenge is to find the appropriate balance between current and future consumption, one that genuinely spreads lifetime income and consumption so as to maintain living standards in retirement. If there were no age pension, the contribution rate would need to be around 18% according to Vince Fitzgerald nearly 30 years ago (as noted by the RIR). Given our means-tested age pension, a typical earner need not save that much but identifying the appropriate SG figure is not a simple issue.
Coates, like the RIR, suggests that the current 9.5% SG will provide more than sufficient income in retirement for the typical retiree. This, however, is based on some questionable assumptions:
- The amount of age pension the typical retiree will be eligible for over their decades of retirement;
- That the income from accumulated superannuation savings should not be indexed to prices (let alone to wages) but should be reduced by the real increases in age pension entitlement that are assumed as the retiree ages;
- An ‘optimal’ pattern of superannuation income that is not, and could not be, offered by superannuation funds.
There are also legitimate questions about the RIR’s modelling, suggesting the likelihood of exaggeration of the likely accumulated savings of the typical retiree (single or couple).
As I have recently argued, it may be that we do not need to increase the SG to as high as 12%, but before we can judge that the Government needs to sort out two key issues:
- The proposed covenant for fund trustees requiring them to offer retirees default products that are in the retirees’ best interests. This will clarify the retirement income products typical retirees should be guided towards (which will not be the RIR’s ‘optimal’product).
- The pension means test so that people can confidently plan their retirement income arrangements, knowing that extra savings will indeed lead to improved retirement incomes, and so that funds can properly identify the products that are in the best interests of their retiree members. There are serious problems with the current test, particularly the assets test.
It would help if Coates and the Grattan Institute focused on these two issues and not simply continue to attack the SG and to suggest unjustified increases in the tax on superannuation.