The equity of government support for retirement income

John, the maximum you could contribute to CSS with employer contribution was 5% of salary. What kind of salary would you have to be on to accumulate $650k? I started work on $9k p.a. Nor is the lump sum available at retirement, just own contributions. It is paid as a form of lifetime annuity.

As David and Kevin Davis note – the desirable shape of the chart is an interesting question.

David prefers a broadly level chart. I believe there is a case for a “tilted V” shape i.e. flat at lower incomes but upwards sloping beyond (say) the 30th income decile.

My reasoning is:

(1) At lower income level, enough support should be provided to ensure no Australian lives in poverty. This will be largely through the age pension.

(2) At higher incomes, the progressive nature of our personal tax system becomes relevant. Higher earners will pay higher amount in tax over their lifetimes hence it is reasonable that the dollar value of tax concessions for retirement income are higher to reflect that. As Ken Henry has said: not every element of the tax & transfer system must individually must meet any particular objective (in this case, progressivity) – it is the overall system that matters. Even with an upward sloping chart, the overall tax and transfer system (including superannuation concessions) is (appropriately) progressive.

Note however I favour caps on the dollar amount of taxation support for very high income earners. Going forward, the contribution caps largely achieve this.

As per usual the politics of envy rears its head with this topic. Lets all just accept that the 20 century has shown that unless you provide incentive to save (including China) you're not likely to do so as see by 70% of retirees with their hands out for the Age Pension. We all need to accept that the 30% not in this cohort aren't going to pay top MTR on investments made outside of super so Treasury's assumption is unrealistic if not absurd. It's remained the assumption simply for political reasons. The question then becomes what's a more optimal long term approach?

Well said.
What are they trying to do? Make everyone poorer? The key is to raise all through productivity, not have endless govt committees restributing wealth like chickens stratching around a pen taking feed from each other.

Given that it is Government expenditure that we are talking about the appropriate discount rate would be the Government borrowing rate, being the bond rate.

Further to my numerical example on 19 March below. The Retirement Income Review says that estimating super tax concessions involves considering the “counterfactual tax benchmark”, which means it compares the tax treatment of contributions and earnings in a super tax setting, with those in a personal income tax system. In other words, it compares the tax paid on super savings with what would have been paid had the savings been in an ordinary savings environment. The difference being the “tax concession”.

In my example therefore, the “concession" in the first year would be $1,460 for someone earning $100k, assuming a personal marginal tax rate of 30% for simplicity (currently 32.5%). The contribution concession being $9,500 x (0.3 - 0.15) = $1,425. The rest being the concession on earnings. Over the 30 years, the earnings tax concession works out to be $1407, undiscounted.

However, this comparison assumes taxpayer behaviour is unchanged. How likely though, would the salary earner be prepared to “save” if favourable tax treatment didn’t exist? And compulsory savings surely would not be implemented without concessional tax treatment. So the "tax concession” is only a concession to the extent that the $9.5k would have instead been taken as salary, and invested outside of super.

In any case, my example shows that the overall tax rate after 30 years on the $9.5k invested, is 30%. The same as the marginal rate that would have applied to the $9.5k if taken as salary and consumed today. So deferring consumption for 30 years leads to the same 30% rate of tax. My point being then, how real are the “tax concessions” in the report?

With the harsh asset & income test, there is virtually no incentive for couples to save over $450,000 in assessable assets. Unlike NZ, which has no income or asset test & there is an actual incentive for retirees to save. All fund managers in Australia should be lobbying the Federal Govt to move to the NZ system.

Thanks SMSF Trustee. Did you understand the question? Let us not compare the size but rather how much risk- free pension is actually received by DB super against the high risk SMSF. So how much pension does Captain Hi Pants receive on his privileged AUD$650k in $$$$. His words not mine. 

Wally, if I've answered a different question to the one you intended, then perhaps that's because of the way you worded it - you asked for a comparison between a $650k Defined Benefit contribution and a $650k SMSF and I said that you can't answer that question without knowing what the defined benefit is. It might pay the same, less or more than the income on an SMSF of that size. You also have to take into account the remuneration that the recipient was paid during their working life as often having a DB scheme is a bit of an offset for a lower salary.

You need more information. I think that actually does answer your question, as best as it can be answered.

Does it help to know that for Federal CSS DB recipients who retired from Commonwealth employment at 55 (as many do but may still work elsewhere), the DB pension was 9.55% of 250% of one's total contributions and accumulated earnings. e.g. Suppose the account held $100k, the retiree is given $100k back tax free, then the scheme assumes $250k remains and the pension (65% of which is reversionary on death) becomes $23,875 p.a. indexed for life (average 30 years) twice a year at the same rate as for aged pensioners. This is almost identical to the single aged pension, without the eligibility hurdle. Because the pension comes from an unfunded source (Future Fund), it is taxable, but with a 10% tax offset. i.e. often no tax at all. That pension level was achievable with less than 10 years service, starting age 27, before moving to the "real world". Little wonder CSS was closed to new members 30 years ago.

John, yes it does help, to some degree. But it leaves out a lot of important information to help someone like Wally who seems to think the scheme is an over-generous rip-off to understand the financials of it.

In particular, CSS is a hybrid scheme, with both parties putting into it. The employee contributes at a lower rate than the Super Guarantee Charge (5% or less according to the CSS website). The employer (ie Government) puts in an additional component. This is in reality similar to someone in an accumulation fund, except the way the contribution is structured is as a payment by the employer out of pre-tax earnings. In both schemes, it's a cost to the employer that is a part of the remuneration paid to the employee.

Thus, the on-going pension payment that you mentioned is a payment linked to the amount that the employer has put in and earnings on that part of the fund. It's also funded by the fact that if someone dies soon after retiring, the fund doesn't have to pay out as much as someone who lives a long time - so it's a scheme that has elements that look a bit like a lifetime deferred annuity structure.

The defined benefit formula had been worked out actuarially based upon longevity and investment return assumptions. Of course, once both of those things changed over the last 20 years or so (life expectancy has lengthened and return expectations have fallen) such schemes needed to close because the formula no longer worked.

But the impression you gave in your response was that this was a huge additional payment for public servants over and above what folk who aren't so lucky get from their super. That is simply not true. It would have been had the scheme been kept open to new members once the investment parameters changed, but the scheme closed when it had to once the actuarial numbers showed that it wasn't sustainable any more.

So, perhaps, no, your comment isn't helpful because it leaves out too much information about how the scheme works and just furthers the misunderstanding of how the public service super schemes have worked over the years. Were they too generous? We still don't have enough information from your response to be able to assess that.

Interesting article. Though it would be useful to know how the super tax concessions are actually calculated. I suspect they may be overstated, being calculated at 15% on contributions and earnings as they arise before and after retirement. Yet if you look at the effect of taxation on contributions and earnings on the end benefit, it appears the tax concessions may not be as favourable as reported.

For example. Someone on $100k today is due to retire in 30 years. Of which, $9,500 is super. After contributions tax, $8,075 goes in, and assume the fund earns 4.50% p.a. The net rate after tax being 3.83%. At the net rate, the $8,075 grows to $24,900 after 30 years.

Whereas if the full $9,500 was to grow at 4.50% for 30 years, the end result would have been $35,600. So the effective tax rate on the end benefit is a full 30%. Double the headline tax rate.

A quick response to Graeme's question

The superannuation tax concessions apply in respect of both contributions (which are paid before retirement) and investment earnings (which are received both before and after retirement) whereas the age pension payments are only received after retirement. Hence, on average, the age pension payments are received later than the super tax concessions.

"For many years there has been considerable media debate about the fairness (or otherwise) of the superannuation tax concessions and the apparent advantage they provide to high income earners ". Just "who" defines fairness? Condemn the high income earners, the most competent, the most successful, the most valuable members of "our" society [...otherwise "we" wouldn't pay them so much, would "we"?] So, what do you want? Lots of low income people earning hand-outs instead? No gratitude for the valued and highly taxed contribution. We all have it so good in Australia! There has NEVER been a better time to be alive but all we get is envy and avarice! They say that "the road to hell is paved with good intentions" but the other side of that is that not all paved roads lead to hell. 

The difficulty I see with David’s paper and the comments thereon are that they address a problem I had hoped had been largely solved for future generations by the Transfer Balance Cap in 2017. It must have been generally recognised at the time of the new legislation that it could not undo all the faults of poor legislation in the past; some perceived “unfair” advantages gained legitimately by a few individuals of the current generation would unavoidably remain as a legacy of the continuing system and allowed to wither on the vine.
Yet David’s paper made no attempt to remove this temporary (but significant) influence of over-large balances when measuring perceived equity between income earner deciles (or asset deciles for that matter) in a mature system. If it had, the analysis would, in my view, provide a sounder basis for indicating what further measures (if any) might be taken.
The 2017 legislation was intended to (and did) reverse the generosity of pre-2007 amendments that had blithely abandoned then-existing maximum benefit limits intended to prevent the accumulation of over-large balances within the superannuation system. We should be wary of taking further precipitate action based solely on a snapshot from an immature system.

The logic and conclusion of your article relies on your statement: "age pension payments are further into the future than the superannuation tax concessions". Your first graph shows that the vast majority of superannuation tax concessions for higher income percentiles are earnings tax concessions - some of which will be received pre-retirement and some (much?) of which will be received post-retirement at the same time in the future as the age pension payments accruing to those in the lower income percentiles.
Can you please justify your statement.

Kevin Davis raises the interesting question as to the desirable shape of this graph. For example, should it be level or downward sloping? I have taken the view that a broadly level graph is preferable to one that is either U-shaped or upward sloping. Of course, that raises the question as to the purpose of retirement income support.

I think there are two reasons:
1 To support the poor
2 To provide some compensation for those who are unable to access their deferred wages due to preservation

Of course the one thing left unconsidered is how the average SMSF individual fares when they have just enough to be ineligible for the pension, but due to falls in interest rates to basically xero, they either have to up their risk profile and invest in things they might not understand the risks of fully and/or live on capital. ( end result not doubt they will become eligible for the pension)
Ask any polititian what was done during COVID for these people and you will find this does not compute as they are the forgotten in betweeners. Meanwhile, those on unemployment , who were sat at home with nothing to spend it on and nowhere to go received double the normal payment. And then there are the Harvey Norman's of this world !!!
the whole thing is an out of control mess.

If they were running an SMSF and didn't understand investments with any capital risk involved then they shouldn't have been running an SMSF in the first place!!!!!

This excellent analysis underlines the need for a National Longevity Strategy for Australians. This would include oversight of decisions relating to longevity by Treasury (Tax and Super) , Social Services (Age Pensions), NDIS, Health and Aged Care, Employment (Age Discrimination and Jobs). We should also include Education: it is a disgrace that we have failed to provide quality longevity education from midlife (compared with what has been done for financial literacy).
Effective integration of these silos through a National Longevity Strategy would go a long way to underpinning our ability to afford the necessary repairs to the aged care system. It would also highlight the importance of making better use of the longevity bonus that continues to accrue to the majority of Australians as they age.

Regardless of the calculations, isn’t the rationale for retirement incomes support policy to assist those unable to adequately finance a reasonable level of living standard in retirement from their own resources. So, rather than focusing on whether the level of government support is relatively equal across the income distribution, the desirable outcome should be that it declines for higher income levels. That is socially optimal policy (sat least from my perspective) should see the graph having much lower height at higher income deciles. Even the most favourable assumptions in the article show that we’re miles away from such an outcome!

You are ignoring the important "one final comment" in the article. This points out that the Treasury analysis makes the incorrect assumption that, without Super concessions, the wealthier folk would be subject to the top marginal rate on their earnings. In practice they would at least focus on generating capital gains which are taxed at half rates and deferred indefinitely until realised. This comes after they first maximize their investment into owner occupied housing which is completely tax free!. This is what we are seeing now. For people making big money now, Super is irrelevant (given the caps on contributions) and they are investing big time in owner-occupied housing. Then they buy investment property.

Thank you for an insightful analysis of the relative levels of taxpayer support for the various income deciles in retirement. Maybe there is another way to look at this. We could look at assets.

The Retirement Income Review noted that there are over 11,000 people, each with more than $5 million in their super fund and each claiming more than $70,000 per year in tax concessions. Indeed, according to one report, the top 100 SMSFs together account for $7.9 billion, or an average of $79 million each. At the same time the ATO reports that the median SMSF is about $700,000. Therefore, 50% of SMSFs are actually smaller than this.

The problem with reporting aggregate super tax concessions, as the Retirement Income Review has done, is that the tax concessions flowing to a small number, but extremely wealthy super funds clearly distort the whole debate about tax concessions flowing to superannuation as a whole.

It may be more meaningful to do this analysis of comparative levels of tax concessions based on retiree’s assets rather than income because funds with such large assets are likely to more concerned with growth than income. Importantly, it would show how these large super funds have been able to benefit from the generous contribution rates that existed before 2007 and are therefore still benefiting from the concessional super tax rates that apply even after the introduction of the Transfer Balance Cap in 2017. It would still show the effects of the Centrelink taper rate because the assets test is the main determinant of the part-pension.

Such an analysis is likely to show that the median SMSF receives very little in tax concessions from super and these retirees also receive very little pension because of the taper rate. It may put an end to the assumption, made by many uninformed commentators, that anyone with a SMSF belongs to the “top end of town”

Great analysis which puts all the moving parts into perspective. In my experience working with pre- retiree clients, the level of knowledge of how the super/retirement income and social security systems work is very low. Those that are prepared to understand it and perhaps seek advice get much better retirement outcomes and in fact have a much greater sense of control and therefore peace of mind

1. What outcome with no assets taper? [Guess: all age pension age eligible better off.]

2. What outcome when (tax paid - welfare benefits) = net tax used? [Guess: low incomers receive, high incomers pay.]

Would like to see how the Defined Benefit Super Schemes of politicians and public serpents compares to your data. They have no skin in the game they want to change but sleep well as their super is linked to CPI increases.

Given that the public sector receive annual wage increases whilst the private sector employees wages are stagnant there is more to this story.

How much super does a politician need in a SMSF to match the million dollar Defined Benefit pensions they currently enjoy.?

Wally, stop living in the past. The public sector super scheme hasn't been defined benefit for years and neither has the politicians' one.

just be aware that with Defined Benefit Super Schemes in pension mode payments over 100K per annum is taxed at 15%
With the Defined Benefit Super Schemes for politicians I do not know if they (the politicians) actually paid for the Defined Benefit Super Scheme but for my own Defined Benefit Super Scheme it cost me $650K AUD for the privilege

Thanks Captain Hi pants....point taken.

To be clear, just how much income does a Defined Benefit Super Scheme of AUD$ 650 k generate ??

Would be great to compare this with a SMSF .

Looking forward to your reply.

Wally, you can't calculate the income from a defined benefit scheme based upon how much the beneficiary has paid in. They pay a defined benefit, usually linked to the final salary or some other measure. They're not market linked. In the case of someone who's paid in $650k, if that's not enough to generate the income needed to pay the defined benefit, then the scheme sponsor (in the public sector case, this is the government) has to supplement the scheme to enable the benefit to be paid.

I know that a private scheme I was once involved in was overfunded and the extent of overfunding was treated by the company as additional reserves on their balance sheet that they could spend however they liked. It didn't belong to the members of the scheme. But, an underfunded scheme is a liability of the sponsoring company, so they have to raise funds to be able to honour the defined benefits.

In response to Wally and Captain Hi pants.

Having previously worked for a state-based DB scheme, the rule-of-thumb was that if a member chose a lump sum at retirement (instead of a lifetime DB pension), then the lifetime rate-of-return they would require on the lump sum to achieve an equivalent outcome to the DB pension is *11% p.a.*

Good luck with that.

It also explains why so many DB schemes were replaced with DC around 30 years ago, and why the Federal Government set up the Future Fund.

Surely the earnings tax concessions enjoyed by the top deciles in Pension Mode will be eroded by the Transfer Balance Cap rules. Understand that there is still a concessional alternative, being the accumulation route, but not to the same extent as the pension route.

Wally, assume the $650k account balance at retirement example used elsewhere in this thread. Extrapolating my experience, the DB benefit for such a Federal public servant CSS participant retiring at age 55 would be: $650k x 2.5 x 9.55% = $158k pa, taxable with a tax offset of 10%, indexed for life, 65% reversionary, plus a tax free (but not levy free!) refund of the $650k, which can optionally be used to generate a second CSS pension on less generous (but better than market) terms. My example is 10 years old, but the rules likely remain for those who joined the APS before CSS closure around 1990. It's feasible for someone joining at age 20 in (say) 1980, working in the APS for 35 years, never rising above middle management.

John I don't know where you get your figures from. But I joined in 1981. My pension is far less than what you calculate, and I was in middle management. And if I die, my estate gets nothing. It is politicians and judges who have the best of it. Also examine the state govt super schemes. These are constitutionally protected funds. The fed govt cannot tax the states, so they escape all fed taxation.

OK, John, but you're still not comparing apples with apples.

Someone with $650k in an SMSF has built that up by contributing the SGC and then it earning income to accumulate that amount.

Someone with $650k as their own contributions plus earnings in the CSS scheme has clearly been on a higher salary in the first place, because they've accumulated the same amount after contributing less than the SGC themselves. This is because the government has put in their contribution on top of the employee's. And it's clear that someone on a higher salary accumulates more than someone on a lower salary!
Further, the imputed amount that's accumulated from the government contribution might look really generous, but that income payment can't be compared with a lump sum permanently invested because it stops when the beneficiary dies (or at least when their spouse dies). So it can afford to be calculated at a rate that is above investment returns because it's more like a deferred lifetime annuity and if the beneficiaries are gone within a year or two of retirement then what they would have been paid is in the pool that gets paid to others.

Another way the pool got boosted was that anyone who joined the public service but left after, say, 6 years, only got paid their contributions and earnings. Whatever the government had put in - at least notionally as most of this wasn't actually funded, but the economics are the same - whatever the government put in during those 6 years is forfeited into the scheme, providing a buffer between any shortfall in investment earnings and the defined benefit payments to those who do stay in it for a long time. DB schemes are designed to reward employees who remain with the employer for a long time.

The average return paid to all beneficiaries is in line with market returns, but each individual can be paid more than that because the pool is grown when someone dies or leaves within a couple of years of commencing.

The scheme closed when the combination of the longevity assumptions and the investment return assumptions got out of line with new actuarial analysis that said people are living longer and investments are returning less.

Look, I'm not an expert in this field. But it really annoys me when people use half-baked partial information to make it sound like the public servants scheme was an overly generous one that someone richly rewarded laziness. It didn't, at least not for most of its existence. Then when it started to be overly generous, it closed to new members. Neither the scheme, nor the public servants who were in it, deserve to be smeared with accusations of excess.

I will add that the 10% offset was a concession made because employees were required to contribute from post-tax dollars, not pre-tax dollars. Nor is there any control over what can be withdrawn each year. The scheme has been closed for many years and most are on modest pensions. You are comparing apples with oranges. Where did you get your $650k from? It is a wild exaggeration.