The proof economy

On the last day of July, the shadow treasurer addressed a room of licensees and product manufacturers at the Financial Services Council’s Shaping Advice Summit in Sydney. He told them the Hayne royal commission had left behind regulatory wreckage, much of it traceable to a single safe harbour provision buried in the Corporations Act.

Chris Dastoor, reporting for Professional Planner, noted in the same piece that a Coalition government received the final report and began implementing it. Tim Wilson’s own website has him overseeing those reforms as chair of the House Economics Committee.

The wreckage is real. Advice now costs more than the households that would benefit most from it can afford. The adviser population has fallen from roughly 28,900 at the end of 2018 to 14,984 in the first week of July, on Padua Wealth Data’s count of the ASIC register.

But both sides of politics have pointed the diagnosis at the wrong object for a decade.

The provision generating the cost is section 961B(2) of the Corporations Act. The industry calls it the safe harbour. Almost nobody outside the compliance function reads what it says.

What the safe harbour provision does

Section 961B(1) states the duty: the provider must act in the best interests of the client in relation to the advice. No qualification, no checklist.

Subsection (2) is different in kind. It opens with a conditional: the provider satisfies the duty in subsection (1) if the provider proves that the provider has done each of the following, and then lists seven steps.

Everything follows from that clause. Subsection (2) is not a description of good advice. It is a rule about evidence, telling an adviser what they must demonstrate after the fact for a regulator, ombudsman or court to deem them compliant. It is a litigation provision wearing the clothes of a professional standard.

If proof establishes compliance, the artefact that matters is the record rather than the recommendation. The record has no natural stopping point. The licensee’s appetite for defensibility, not the client’s need, determines its size.

This is also why scope will not sever. A client who walks in with one question still generates a file establishing the whole position. That file has to survive the later argument that the question could not properly have been answered in isolation.

Cost becomes fixed rather than variable. Fixed cost sets a minimum viable fee. That minimum viable fee is the advice gap expressed as arithmetic.

Why intent did not survive the structure

None of this was intended. The revised explanatory memorandum said the steps were not meant to be an exhaustive and mechanical checklist. It also said an adviser could comply with subsection (1) without recourse to subsection (2). The note that remains in the Act anticipates scaled advice and tailored inquiries.

Structure, not culture, defeated those assurances. A deemed-compliance pathway is only nominally optional. Once it exists, indemnity underwriters price the alternative and licensees mandate the pathway to protect the licence. Satisfying subsection (1) directly then becomes commercially unavailable to anyone who wants cover.

The Financial Services Council said as much in its 2020 submission to ASIC’s Consultation Paper 332. It observed that codification had turned the steps into an entrenched seven-step process and the only practical route to compliance. Satisfying them, it added, does not mean the advice was in the client’s interests.

Where the steps lead

Step (e) is the one that matters. Where it would be reasonable to consider recommending a financial product, the adviser must conduct a reasonable investigation into the products that might achieve the client’s objectives and meet their needs.

Commissioner Hayne found that this required advisers to make little or no independent inquiry into or assessment of products, because in practice the products came from the licensee’s approved product list. The step meant to guarantee diligence instead ratified whatever the licensee had already decided to distribute.

So the framework is expensive, and it is expensive at documenting distribution. The cost is not the price of professional judgement. It is the price of evidencing a procedure whose operative step terminates in product selection.

Hayne recommended repeal in Recommendation 2.3, subject to a review showing clear justification for retention. The review was to report by the end of 2022.

Every intervention since has failed to move it. Michelle Levy’s Quality of Advice Review recommended removal. Stephen Jones announced it in December 2024. Treasury then dropped the measure from the Delivering Better Financial Outcomes tranche two exposure draft in March 2025.

When Daniel Mulino faced the question in July of whether he remained committed to the principle, he said he would be careful about how much he committed to.

14 years on, Parliament has left the subsection unamended.

Why we regulate procedure

We regulate the procedure of advice in exhaustive detail because we never settled a knowledge standard against which to assess judgement.

The obvious objection is that standards exist. Section 921B (2) sets a qualifications standard. The Corporations (Relevant Providers Degrees, Qualifications and Courses Standard) Determination 2021 lists approved degrees. Treasury’s 2022 consultation paper records the 11 knowledge areas an approved degree must cover, with an exam, a professional year and a code of ethics on top.

All real, and none of it does the work in question. Those instruments govern who may be authorised, accrediting a curriculum and a person at the point of entry. What they do not supply is any means of assessing a particular recommendation.

Nothing states what a competent practitioner must be able to determine about a given household’s position. No benchmark exists for testing the judgement in a file.

The experienced provider pathway sharpens it: an adviser with 10 years up to the end of 2021 and a clean record could meet the standard without the degree at all.

The file substitutes for the standard. That is not a drafting failure. It is what happens when the people who built the framework conceived the activity as distribution rather than as knowledge work.

The Compensation Scheme of Last Resort is where this surfaces as a number. Its revised estimate, released on 2 July, puts the 2026-27 cost at $198 million. Of that, $190.3 million is attributed to personal financial advice, up from an initial November estimate of $126.9 million on the strength of the final Dixon Advisory cohort and the first tranche of Shield and First Guardian claims. Two product failures, and the levy lands on advice.

The folk model, rendered as statute

The shadow treasurer said something more revealing than the line about wreckage. He wants fee structures “aligned with incentives and success”, and the success he named was best returns.

That is what advice looks like from the outside. Ask someone who has never engaged an adviser what an adviser does. The answer will be some version of picking investments that go up. It is not a foolish belief; it is the only visible part of the job.

It is also wrong in a specific way. Markets produce return. What the adviser controls is the structure of the balance sheet, the tax treatment, the contribution and drawdown pattern. The adviser also determines whether cover is in place before it is needed, and whether a household can absorb a shock without a forced sale at the worst moment.

Those decisions determine what the client realises. That is a money-weighted outcome rather than an index number. Two households in the same portfolio across the same decade can finish in materially different positions. The gap is the advice.

What the performance test reveals

The Your Future, Your Super performance test has run annually since APRA introduced it in 2021. It measures a product’s net investment return over eight years against a benchmark built from that product’s own strategic asset allocation. It fails the product if its return falls more than half a percentage point behind.

Two consecutive failures and the product closes to new members. The predictable response is to hold the benchmark. Treasury’s own consultation paper on reforming the test, released this year, records the concern without euphemism. The test encourages benchmark hugging. Trustees invest more closely in line with the benchmark indices to manage the risk of failure.

So one arm of government compresses the dispersion of superannuation returns by design, while a prospective treasurer proposes to pay advisers according to it.

The compression is also uneven. The test does not reach retirement phase products. It has not reached externally directed products either, which is precisely where Shield and First Guardian sat.

Treasury draws the connection itself. The collapse of two schemes distributed through superannuation platforms has drawn attention to that exclusion. The test constrains return where it measures and leaves it unmeasured where the money went.

The point is not that a politician holds the lay view. It is that the lay view is the design basis of the statute. If advice is fundamentally product selection, a duty about advice is naturally a duty about how the adviser selected the product. Its operative step naturally terminates in an investigation of financial products. Paragraph 961B(2)(e) is the folk model written into law.

The framework is expensive at documenting distribution because the people who built it believed distribution was the thing being done.

What would work

Three changes, and they only work together.

Repeal the process test. Section 961B(1) can stand alone. The duty was never the problem; the deemed-compliance pathway was.

Build the assessment standard that entry requirements were never designed to be. Not another exam, and not a longer list of approved degrees.

A body of technical knowledge detailed enough that a court or ombudsman can assess judgement against it, rather than counting steps.

Without it, repeal relocates the uncertainty to insurers and licensees, who will rebuild the checklist privately and less coherently than Parliament did.

Move the duty upstream, and do not simply copy the United Kingdom. The design and distribution obligations in Part 7.8A require an issuer to determine a target market, set distribution conditions, take reasonable steps toward consistent distribution, review and notify ASIC of significant dealings. Every step asks whether a product reached the right consumers. None asks whether it was worth owning.

The Financial Conduct Authority’s Consumer Duty, in force from 31 July 2023 for open products, is the obvious reference point.

Two features are worth taking. Principle 12 binds any firm that determines or materially influences retail customer outcomes across the distribution chain, including manufacturers with no retail customers of their own.

The handbook states plainly that it creates no fiduciary relationship and requires no firm to give advice. That disposes of the objection that a manufacturer duty is merely the best interests duty in another coat.

What should not be taken is the enforcement architecture. Regulators judge a duty to deliver good outcomes after the event, against a standard nobody can state in advance. British firms now produce annual board-approved outcomes assessments in response.

The supervisory question has become whether a firm can prove its customers received good outcomes. That is the proof economy again, one level upstream.

Falsifiable rather than aspirational

The better design is falsifiable rather than aspirational. Require a manufacturer, before issue, to state the envelope.

That means the conditions under which the product will fail its holder, the liquidity it can deliver under stress as distinct from the redemption terms it advertises and the range of paths a target holder should expect rather than a terminal expected return. Then attach liability where realised experience falls outside that envelope for reasons inherent to the design.

An outcomes duty asks whether the firm tried hard enough. Documentation answers it. An envelope duty asks whether what the manufacturer said was true. Comparing a statement to a fact answers it. One generates files. The other generates a testable claim.

Two things follow. Path becomes the unit of assessment rather than terminal expectation. That is the correct unit for anyone drawing down a superannuation balance, where the order of returns determines the result and the average conceals it.

Loss attribution can follow the failure. Put manufacturers and trustees into the compensation chain. The party that designed the failure should price it at origination rather than socialise the cost across an advice sub-sector invoiced $190.3 million for two collapses it did not build.

Do one of the three and the system gets worse. Do all three and advice becomes what it was always supposed to be: a knowledge profession earning its price through the judgement it supplies rather than the paperwork it must survive.