An honest guide to running a self-managed super fund

The SMSF industry has a built-in tilt toward optimism about self-management.

Advisers who specialise in them, accountants who service them, and lawyers who draft the deeds all have structural incentives to recommend one. That is worth saying plainly, including of anyone writing a guide like this.

An SMSF is the right answer for some people, in some circumstances, with a particular level of commitment. It is emphatically not the right answer for everyone, and the cost of getting the decision wrong — establishing one when a well-run industry fund would have served you better, or establishing one without adequate support — is measured in six figures often enough to take seriously.

What follows is the framework, the rules that cannot be bent, and the two things that changed this year.

What an SMSF actually is

Three identities stacked in one structure: a trust, a superannuation fund, and a fund you run yourself.

To be a validly regulated fund it must satisfy four conditions.

Membership cannot exceed six, and most funds have between one and four members — typically a couple, or a couple plus adult children. Every member must also be a trustee, or a director of the corporate trustee; you cannot be a member without being a trustee, which is the "self-managed" part made concrete. The fund must satisfy the sole purpose test. And it must comply with the investment rules of the SIS Act.

On structure: a corporate trustee — a private company established to act as trustee — is worth the modest extra cost in nearly every case. Adding or removing a member does not require re-registering every asset, there is a further layer of limited liability, and the resulting fund is simply cleaner.

The cost question, answered honestly

Retail and industry fund fees are a percentage. SMSF operating costs are largely fixed. That difference creates a breakeven point, and it is the whole of the financial argument at the lower end.

The commonly cited threshold is somewhere around $200,000 to $300,000. In practice, waiting until $300,000 to $400,000 is the more defensible position. At $200,000, annual operating costs of roughly $4,500 are a 2.25% drag — worse than almost any retail fund. At $400,000 the same cost is 1.125%. At $600,000 it is 0.75%, already competitive. Above $1,000,000 a fixed cost beats a percentage every time.

But the fee comparison is not the case for an SMSF. It is only the case against one below a certain size.

The real argument is access: to assets and strategies a pooled fund cannot offer you — direct commercial property, a genuinely concentrated position held deliberately, control over the timing of every capital gain. If none of those matter to you, the fee saving alone is a thin reason to take on the obligations that follow.

Four questions worth answering before you sign anything

Is the balance large enough? See above. The exception is a strategy whose returns are high enough that the fee comparison stops mattering — but "I expect higher returns" is a belief, not a plan, and it should be written down before it is relied upon.

Do you have access to real professional support? Not a discount administration service. Someone who will tell you no.

Are you willing to be genuinely engaged? An investment strategy reviewed annually, minutes kept, an audit each year, decisions documented before they are made rather than justified afterwards.

Do you understand the sole purpose test? If the honest answer is that you intend to keep a holiday apartment in the fund and use it occasionally, stop here. The next section explains why.

The rules that do not bend

Six provisions cause almost all serious SMSF failures. In nearly every case the trustee believed they understood them.

The sole purpose test (s 62). The fund must be maintained solely to provide retirement benefits to members, or death benefits to their beneficiaries. Solely. A trustee who used a fund-owned beachfront apartment for a few weeks a year — around eleven percent of its available use — did not make the fund eleven percent non-compliant. The fund was non-compliant, full stop. Partial personal use is not a partial breach.

The related party asset prohibition (s 66). An SMSF cannot acquire assets from a related party, and "related party" is drawn broadly: members, their relatives, and the companies and trusts they control. Three narrow exceptions exist — listed securities at market value, business real property at market value, and in-house assets acquired within the 5% limit. Residential property you already own does not qualify, however commercial the arrangement is made to look.

The lending prohibition (s 65). The fund cannot lend to a member or a member's relative, or provide financial assistance to them. "Financial assistance" reaches past formal loans into guarantees, indemnities, and any arrangement that shifts risk off a family member. Helping an adult child start a business from the fund is one of the most expensive kind gestures available in Australian law.

The investment strategy requirement (s 52B). The strategy is a legal document, not a wish list. It must address risk, return, diversification, liquidity and the insurance needs of members — and it must exist before the investment, not be reverse-engineered after it.

The in-house asset rule (s 71). Widely misunderstood. There is no general five percent cap on any single investment; a fund may lawfully hold eighty or ninety percent of its assets in one commercial property, provided the strategy properly addresses diversification and liquidity. The five percent cap applies to in-house assets, which is a defined term — broadly, loans to or investments in related parties.

The collectibles rules (reg 13.18AA). Art, jewellery, antiques, coins, wine, classic cars, boats. Not prohibited, but hedged with conditions on storage, insurance, use and valuation that exist because these assets share one property: the temptation to enjoy them. A classic car that is driven is no longer an investment.

The tax architecture

An SMSF lives in one of three tax worlds, and the distance between them is the reason the structure exists at all.

Accumulation phase taxes earnings at fifteen percent. Retirement phase — assets supporting a pension, within the transfer balance cap — taxes them at nothing. That is the destination, not a bonus.

The third world is the one nobody plans for. Non-arm's length income and expenditure — NALI and NALE — taxes the affected income at forty-five percent. It is triggered by dealings that are not on commercial terms, including expenses that are too low: a discounted service from a member's own firm can taint the income it relates to. NALI is not a penalty that can be paid off and forgotten. It attaches to the income, permanently.

Franking credits deserve a mention because they behave unusually in this structure. A fund in retirement phase paying no tax does not merely receive the credit as an offset — it receives it as a refund. Over a long retirement, on a portfolio of fully franked Australian equities, that mechanism does a great deal of quiet work.

What changed in 2026

Two changes land this year, and both post-date most of the SMSF material in circulation.

Borrowing to buy residential property is over

The Treasury Laws Amendment (Tax Reform No. 1) Act 2026 received Royal Assent on 26 June 2026, and the relevant provisions commenced on 10 August 2026.

From that date, where an SMSF uses a limited recourse borrowing arrangement to acquire real property, the property must be business real property. Residential property is excluded — unless it is used wholly and exclusively in one or more businesses, which is a narrow gate. A mixed-use building with an office below and an apartment above does not pass; partial residential use disqualifies the whole asset.

Three things are worth being precise about, because the change is narrower than "SMSFs can no longer borrow".

Existing arrangements are grandfathered. A residential LRBA entered into before commencement remains valid, and refinancing an existing borrowing is still permitted. Contracts exchanged before 10 August 2026 are protected even where settlement happens afterwards. And LRBAs over non-real-property assets — listed shares, units in widely held trusts — are untouched.

What has genuinely closed is the strategy that drove a great many SMSF establishments over the past decade: set up a fund, borrow, buy a residential investment property inside it. That door is shut for new arrangements. If an SMSF was being established primarily to execute it, the reason for the fund has gone, and the honest response is to revisit the structure rather than to look for a workaround.

For business owners, the more useful half of the rule survives intact. Buying your own commercial premises through the fund and leasing it back to your business at market rent remains available, and remains one of the few genuinely elegant strategies in Australian superannuation.

Division 296 applies for the first time this year

The additional tax on large balances takes effect from 1 July 2026, which makes 2026-27 its first year of operation.

The mechanics, as legislated: an additional fifteen percent applies to earnings attributable to the part of your total superannuation balance above $3 million, and a further ten percent — twenty-five in total — above $10 million. The thresholds are indexed.

Two details matter more than the headline rate.

The final legislation taxes realised earnings only. Unrealised gains were removed during its passage, and a notional cost base set at market value on 30 June 2026 is tracked so that gains accrued before commencement are not caught. This is a material improvement on what was originally proposed, and it changes the planning question from "how do I fund a tax on paper gains" to the ordinary one of when to realise.

The tax is assessed to the individual, not the fund — though the fund can be directed to pay it. For SMSF trustees the practical consequence is reporting and valuation: asset values at 30 June now carry a consequence they did not previously carry.

For 2026-27 only, the balance is measured at 30 June 2027. In later years it is tested at either end of the year, whichever is higher.

Figures in this section are current as at August 2026 and are indexed or amended from time to time — check them against ATO guidance before acting.

The parts people leave until it is too late

Insurance. The investment strategy must address whether the fund holds insurance for its members, and a great many strategies dispose of this in a single sentence. Leaving a pooled fund usually means leaving default cover behind — cover that was underwritten without medical questions and cannot always be replaced on the same terms once you are older, or once something has been diagnosed. Check what you are giving up before you roll out, not after.

Death benefits. Superannuation does not pass under your will. It passes according to the trust deed and the fund's rules, and in an SMSF the person exercising the discretion may be the surviving trustee — who may be a second spouse, or an adult child with an interest of their own. A binding death benefit nomination, kept current and drafted to the deed's requirements, is the difference between an intention and an outcome.

There is a tax dimension people rarely see coming. Adult children who are not financial dependants pay tax on the taxable component of a death benefit. On a large balance this is a meaningful sum, and there are legitimate strategies — recontribution over time, reversionary pensions to a spouse, timing withdrawals in the final period of life — that reduce it. All of them require time, which is exactly what nobody has when it becomes urgent.

The exit. Funds wind up for ordinary reasons: age, illness, a relationship ending, or moving overseas. That last one carries a trap — an SMSF whose central management and control moves offshore for too long can fail the residency test, and the consequence is taxation of the entire fund at the top marginal rate. If you are leaving Australia for an extended period, the structure needs to be dealt with before you go, not remembered from abroad.

What actually separates the good outcomes from the bad

Not intelligence. The cases that end badly are full of accountants, engineers and executives who read contracts for a living.

What separates them is whether there was someone in the structure whose job was to say no, and whether the trustee was willing to hear it. In most of the expensive failures, either nobody was asked, or somebody was asked and overruled.

An SMSF gives you control. Control is not the same as advantage — it is only the opportunity to have one, and it comes attached to obligations that do not care whether you were busy that year.

The best fund is not the one with the cleverest strategy. It is the one you can actually run.