The member decision gap

Welcome to TRR #9 for 2026. This edition arrives with the backdrop of escalating Middle East tensions still reverberating through global markets. A reminder, if one were needed, that retirement savers and trustees alike are never entirely insulated from the world beyond superannuation policy.

We cover ASIC's landmark national survey on retirement readiness (the numbers are confronting), the wave of investment switching behaviour triggered by geopolitical anxiety, and APRA's long-awaited shot across the bow on AI governance. Plenty to chew on.

Australians approaching retirement are anxious, underprepared and, perhaps most worrying, largely unaware of the gap between their confidence and their actual financial literacy. That is the uncomfortable finding at the heart of a major national survey released by ASIC during April, undertaken in conjunction with the launch of its Moneysmart Retirement Hub.

The research, which surveyed more than 2,000 Australians aged 45 to 75, paints a picture of Australians knowing that they aren’t prepared for retirement, but struggling to close both planning and knowledge gaps. Of the pre-retirees surveyed (those aged between 50 and 66), the following stand out:

The confidence gap is particularly striking. While 41% of pre-retirees say they feel confident managing their retirement finances, only 26% demonstrate a genuinely strong understanding of retirement finance concepts. Nearly half (46%) report low financial literacy and low confidence in tandem. In other words, many Australians who feel retirement-ready are, by objective measures, far from it.

Cost of living is amplifying the anxiety. Some 55% of pre-retirees expressed concern about the impact of rising living costs on their retirement plans. The research also highlights a stark gendered dimension: female pre-retirees report markedly lower retirement confidence, with just 33% feeling confident about managing their finances compared to 43% of male pre-retirees. Renters fare worst of all, with only 16% confident about living comfortably in retirement.

ASIC contends that irrespective of financial position or confidence, five questions tend to dominate the typical pre-retiree’s concerns:

These numbers should be required reading for every trustee board in Australia. The fact that only one in five pre-retirees has a clear retirement plan is not primarily a consumer problem, it is a system design problem. For decades, superannuation has done an excellent job of accumulating assets for members, and a markedly less impressive job of helping them understand what to do when the accumulation phase ends.

The confidence-literacy gap is particularly revealing. Members who feel confident but are not financially literate are precisely the cohort most at risk of making poor decisions at the point of retirement, and the least likely to seek advice because they believe they don't need it. Trustees would do well to examine whether their own member communications are bridging this gap or inadvertently reinforcing it.

Better practice in supporting member engagement with retirement income solutions, particularly Principles 12 to 16 of the Best Practice Principles, would go a long way toward addressing many of the issues raised in ASIC’s research.

ASIC's Moneysmart tools are a welcome and credible public resource, but they cannot substitute for personalised guidance and advice, delivered at scale. The best-informed member is one who has received tailored communication from their fund, not just a government website. The question for trustees is whether their retirement income strategies are genuinely closing this gap, or merely gesturing at it.

The escalation of the Middle East conflict in March, including US and Israeli strikes on Iran and the subsequent closure of the Strait of Hormuz , did what geopolitical shocks reliably do: it sent oil prices surging, equity markets lurching, and a wave of anxious superannuation members scrambling for the safety of cash.

Two of Australia's largest industry funds, HESTA and UniSuper, have broken from their usual reserve to speak publicly about what they observed, and both should be commended for doing so. Their insights serve as a timely reminder of the constant liquidity tension funds face between managing long-term assets for members and their reflexive tendency to ‘dash to cash’ in every significant market downturn.

HESTA: The $20,000 lesson from COVID

The numbers during March tell a clear story at HESTA. Visits to the fund's investment landing page climbed nearly 37%, and average daily switching activity rose compared to the prior month. The peak came on 9 March, when oil prices broke through US$110 per barrel and the ASX 200 fell 2.8%, its worst single day since the tariff shock of April 2025. According to HESTA, members who switched moved predominantly into very defensive options, primarily Cash and Term Deposits.

Outgoing HESTA CEO Debby Blakey used the event to invoke a pointed piece of history: a member with a $100,000 balance who switched from HESTA’s default Balanced Growth option into Cash and Term Deposits during the COVID market shock in 2020, and who took a year to switch back, could have been more than $20,000 worse off five years later. The cost of what behavioural economists call ‘myopic loss aversion’ is, in other words, far from abstract.

UniSuper: Inflation, not war, is the real risk

UniSuper's CIO John Pearce offered a more structural read of the situation. In a recent investment update, Pearce observed that the geopolitical shock had produced a relatively short-lived equity sell-off, with investors quickly pivoting to focus on the downstream question: would higher energy prices prove inflationary enough to derail corporate earnings and force central banks to act?

Pearce also noted that UniSuper members had shifted around $1.4 billion from growth options into more defensive settings during the recent volatility, a significant figure that underscores the behavioural challenge facing trustees whenever markets dislocate.

His closing observation echoed a familiar truth: cash is the most reliable short-term haven during acute market stress, but moving to the sidelines carries a significant long-term cost if investors miss the strongest recovery days. "Time in the market beats timing the market" may be an old line, but the data consistently validates it.

The switching surge throughout March would not have been experienced only at HESTA and UniSuper. It’s a system-wide behavioural dilemma, and a challenge that sits squarely at the intersection of product design, member communication, and retirement income strategy.

For accumulation-phase members, panic switching is damaging. For members in or approaching retirement, it can be catastrophic: crystallising losses, disrupting sequencing risk management, as we wrote about recently, and potentially triggering tax consequences for near-retirees.

What's striking about both funds' public responses is how explicit they are in providing the counterfactual data; not just the message to stay calm, but the dollars-and-cents evidence of what switching has historically cost. This is good communication design.

But the deeper structural question is whether the product architecture itself is doing enough work. Lifecycle options with automatic de-risking glide paths, or retirement income products that smooth volatility rather than exposing members to raw market moves, are precisely the kind of design responses that reduce the frequency and severity of these behavioural events.

In short, the best defence a trustee can have against panic switching is a product that makes switching less likely, even absent any comms.

Trustees who are still relying primarily on after-the-fact communications, issuing "don't panic" messages after members are already triggered, may want to ask whether their retirement product suite is doing enough of the heavy lifting before the next significant market drawdown arrives. Because that isn’t a question of ‘if’ but ‘when’.

Superannuation trustees have had AI on their radar for some time, using it in investment analysis, member communications, fraud detection, and increasingly in customer-facing applications. APRA is now signaling that the governance frameworks surrounding these deployments have not kept pace with the technology itself.

The regulator recently published a letter to all regulated entities: banks, insurers, and superannuation trustees alike, calling for a material step-change in how AI risks are identified, managed, and overseen.

The letter follows a targeted supervisory review APRA conducted late last year across its regulated industries, examining both how AI is being deployed and how it is being governed. The findings are instructive, and in some places uncomfortable.

Entities are moving from experimentation toward operationally embedded and customer-facing AI applications, but governance arrangements have not matured commensurately. Boards demonstrate strong interest in AI's potential benefits but, according to APRA, many lack the technical literacy required to provide effective challenge to management on AI-related risks. That is a significant gap in the context of the SPS 510 director competency obligations that already apply to trustees.

Concentration risk is flagged explicitly: some entities are heavily dependent on a single AI provider for multiple use cases, with insufficient contingency planning. APRA also points to the opacity problem; AI functionality increasingly embedded within broader software platforms or developer tooling, reducing transparency over where and how models are trained or updated, and limiting entities' ability to fully assess the risks they are assuming.

Perhaps most pointed is APRA's observation about the threat landscape. The regulator warns that frontier AI models are expected to increase both the probability and scale of cyber attacks, as sophisticated tools lower the barrier for malicious actors to discover and exploit vulnerabilities. APRA Member Therese McCarthy Hockey framed the challenge directly: the speed at which entities can identify and patch vulnerabilities needs to operate much faster, commensurate with an AI-accelerated threat environment.

Notably, APRA stopped short of proposing new prudential standards. The expectation is for significant improvement under existing frameworks, particularly CPS/SPS 234 (information security), CPS 230 (operational risk), and governance standards rather than the creation of new AI-specific requirements, at least for now.

APRA has been measured and deliberate in how it has approached AI regulation, far more so than some of its counterparts in other jurisdictions. The decision not to introduce new standards at this stage, and instead to set an expectation of improvement under existing frameworks, reflects a pragmatic judgement: the technology is moving too quickly for durable new rules, and the bigger near-term risk is complacency rather than over-prescription.

For superannuation trustees, the board competency observation deserves particular attention. APRA is essentially signaling that it expects trustee boards to be capable of genuinely interrogating management's AI risk frameworks, not just receiving briefings and nodding uncritically. Given that many boards are still building baseline digital literacy, this is an area where trustee education investment will have direct regulatory dividend.

The concentration risk finding also has retirement income implications that go beyond the obvious operational concern. If a fund's member engagement, advice delivery, or retirement income modelling is heavily reliant on a single AI provider, and that provider experiences a material disruption or undergoes a significant model update, the downstream effect on member experience and potentially on member decisions, could be substantial. Diversification of AI dependencies is not just a technology architecture question; it is a member outcomes question.

Trustees who have not yet mapped their AI use cases against APRA's existing prudential standards, and assessed where the gaps are, should treat this letter as a prompt to do so. The regulator has telegraphed its expectations clearly. The next supervisory engagement cycle will likely test whether the industry has listened.