Only a few months ago, it seemed this new tax was almost certain to become law (“this new tax” being the highly controversial extra 15% tax on super earnings for those with more than $3 million in super – aka Division 296 tax). At the time, a newly re-elected and confident government was firmly sticking with its intention to reintroduce the bill and get it passed. It seems amazing that we’re now speculating about changes. But here we are.
In recent weeks it seems even the Treasurer’s resolve has wobbled and there is a quiet murmuring of potential change. If the rumours are to be believed, the two aspects of this legislation being reviewed are the two most contentious: the lack of indexation of the $3 million threshold and the taxation of unrealised capital gains.
It’s likely the Treasurer hopes throwing a bone like indexation will take our eyes off the bigger issue of taxing unrealised gains.
It shouldn’t.
Indexation would be nice – and definitely better than nothing - but nowhere near as powerful as we might like to think.
Firstly, indexation would largely benefit those not yet caught by Division 296 tax. For example, if the threshold is indexed to inflation and this is around 3% pa, the threshold would reach $4 million in around 10 years, $5 million in around 18 years and $6 million in around 24 years. This is excellent for someone starting with $1 million today and maximising their super over the next 20 years. Whereas they might be likely to exceed $3 million, the chance of exceeding $5 million - $6 million in that timeframe is far lower. This is because the contribution caps we already have will quickly limit their ability to boost their super with their own money.
For those already in Division 296 tax territory, indexation would definitely result in less tax but is unlikely to have the enormous impact many would assume.
Consider the following example:
- Bob has $7 million in super initially (including a $2 million pension) – minimum pension payments are drawn each year
- The $3 million Division 296 tax threshold increases in line with inflation (and let’s say that’s 3% pa)
- The Fund’s investment return is 8% pa before tax (5% growth, 3% income)
After 10 years, Bob would be around $40,000 wealthier if the $3 million threshold is indexed vs if it remains fixed at $3 million (this has been adjusted for inflation – so it’s $40,000 in “today’s dollars”).
That is certainly a valuable saving. But over that time, how much Division 296 tax has he paid? In this example, he’s paid around $670,000 in Division 296 tax in total even if the threshold has been indexed.
Repeating the same exercise with someone who started with $10 million in super, the saving thanks to indexation is roughly the same (around $40,000). But this time, the Division 296 tax paid during that time is over $1.1 million even if the threshold is indexed.
Why is it having so little impact?
This isn’t the right maths but conceptually, a $100,000 increase in the threshold means the member avoids 15% tax on earnings on an extra $100,000. If earnings (both income and growth) equate to 8% pa (for example), all this person is really saving is 15% x 8% x $100,000 (around $1,200). Of course this compounds over several years and the difference grows each year (ie, the gap between the indexed threshold and $3 million gets bigger) – so after 10 years, it’s a meaningful amount. But it’s around the same saving for everyone (no matter how big their starting balance) and it’s therefore dwarfed by the amount of Division 296 tax the member has actually paid for larger balances.
What does this mean?
Those fighting for change when it comes to Division 296 shouldn’t settle for indexation of the threshold. It’s better than nothing but doesn’t really change the equation for larger balances. And it doesn’t solve the fundamental issue of taxing unrealised gains.
That problem has been well articulated by many people in this debate in that the current structure results in:
- highly unpredictable tax bills for those with volatile assets – and therefore a tax cost that simply cannot be planned for,
- tax being imposed at a time when there is no guarantee there will be money available to pay it, and
- tax being paid on asset growth that subsequently disappears.
Interestingly, my modelling suggests that if the Government found a way to tax actual capital gains when realised (ie the usual approach when it comes to taxing things), they might even raise more revenue.
In particular, without some form of grandfathering, this approach would effectively result in an additional tax being applied to gains that have already accrued (before the introduction of the new tax).
For example, let’s imagine Bob’s $7 million super fund has already built up $2 million in accrued (so far unrealised) capital gains at 30 June 2025.
Under the current approach, the ‘earnings’ taken into account for Division 296 tax would be his fund’s income (3% less tax) plus its growth (5%). Let’s say that amounts to around $540,000. So if Division 296 tax was introduced exactly as planned, he would pay 15% tax on a proportion of this amount (in his case, the proportion happens to be around 60%).
But what if the method switched to a proportion of ‘actual taxable income’ (rather than the earnings amount above). If Bob’s SMSF didn’t sell any assets during the year, his earnings would only be around $210,000 (3% of $7 million).
However, if his fund sold some of its assets and realised (say) $900,000 in capital gains, the Division 296 earnings (on an ‘actual taxable income’ method) might be more like $810,000. This is the fund’s income of $210,000 plus his fund’s capital gains, discounted by one-third (this is the discount super funds normally get for assets held for more than 12 months) - ie $600,000. Again, we can assume Bob would only pay Division 296 tax on a proportion of this amount (to reflect the fact that not all of his super is over $3 million).
This is because if the method was changed and no grandfathering was provided, even the gains Bob’s SMSF has built up before Division 296 tax would be caught. In contrast, Division 296 tax as it stands right now only looks forward – only growth post its introduction is taxed.
Of course, any change in method would need to consider:
- whether it’s appropriate to consider grandfathering to avoid retrospectively taxing gains that have already built up over many years,
- how to allow for discounting of capital gains (Division 296 tax as it stands effectively ignores discounting – it simply captures ‘growth’),
- how to deal with pension accounts (in my examples above I’ve assumed the tax would be applied to a proportion of income ignoring any reduction due to the fact that some of the fund’s normal taxable income is entirely tax exempt because of its pension).
The key, however, is that the government probably could find a way of capturing just as much tax revenue even if it changed the tax. The problem is it would have to wait longer for it to arrive. If Division 296 tax is aligned to actual taxable income, the big tax bills would only really emerge when capital gains were realised.
But at least the people subject to the tax could be guaranteed to have the cash when it happened.
Meg Heffron is the Managing Director of Heffron SMSF Solutions, a sponsor of Firstlinks. This is general information only and it does not constitute any recommendation or advice. It does not consider any personal circumstances and is based on an understanding of relevant rules and legislation at the time of writing.
New: The Heffron SMSF 2025/26 Facts and Figures document has been finalised and is available as a free download. Keep it on-hand to access the most recent information to stay up to date.
For more articles and papers from Heffron, please click here.