Australia’s private credit sector was already under regulatory scrutiny after a run of fund collapses and redemption freezes. The latest episode was different: a dispute between a manager and its auditor over what the loans were worth.

In September 2026, KPMG declined to issue its audit opinion on the FY2026 financial statements of three Metrics vehicles by the statutory reporting deadline: Metrics Master Income Trust (MXT), Metrics Income Opportunities Trust (MOT) and Metrics Real Estate Multi-Strategy Fund (MRE). The disagreement concerned how private-market assets were measured, including expected credit loss provisions, fair values and valuation assumptions. Metrics had relied on Charter Keck Cramer for an initial property valuation and on Alvarez & Marsal, appointed by the funds’ trustees, for a valuation range on the loans. The dispute was resolved in early October, when KPMG signed the accounts after roughly A$170 million of write-downs across the three trusts. These included higher expected credit loss provisions, more weight on downside scenarios and lower fair values for some assets.

It is tempting to read this as an ordinary audit disagreement. We think it says something more useful about how private asset valuations work. The machinery around a private valuation, from independent ranges to smoothed inputs to process-focused assurance, is built to absorb change. Marks move when outside pressure makes holding them costly. At Metrics, that pressure landed on the auditor.

This post draws on fourteen interviews with valuation professionals at superannuation funds, insurers, asset managers (GPs and LPs) and pension fiduciaries in Australia and Europe, together with our international survey of valuation practitioners and a review of seven regulatory regimes.

Why private marks stay put

A range sets limits; it doesn’t test the number

Metrics’ structure, with an independent firm supplying a valuation range and the manager choosing a figure within it, is common. What the Charter Keck Cramer and Alvarez & Marsal engagements covered has not been made public, and all three firms declined to comment. The general model is still worth examining.

Several interviewees described the same process. The manager’s internal valuation is checked against the valuer’s range, and the aim is to confirm it falls inside. One told us it always has. That limits how far the reported value can drift, but it says little about whether the value is right.

The range held even when an investor pushed back. The head of valuations at a large superannuation fund told us his team spotted a conflict of interest in how a manager and its valuer had reached a figure and adjusted it. The adjusted figure still sat inside the valuer’s range. For distressed or illiquid loans, those ranges can be wide, leaving plenty of room for the manager’s judgement.

Assurance checks the process

No one we spoke to could point to a case where an external review forced a valuation to change. At one pension fiduciary, a valuation passes through an external adviser, an in-house team, the auditor’s own valuation specialist and an audit committee. The officer responsible, himself a former audit partner, could not recall any of them prompting an adjustment. He called the exercise verification. At another manager, the valuation function holds a formal veto over the published figure and has never used it.

Internal challenge does change numbers. Reviewers catch stale comparables, inconsistencies and arithmetic errors. External assurance does something narrower, it confirms the process was followed and the stated inputs were used. Whether the resulting figure is right is a question it rarely reaches.

Inputs are smoothed and news is netted off

A higher cash rate and changes to the tax treatment of property investment both lower collateral values and raise discount rates and default probabilities. Listed prices reflect that within days. In a private portfolio, it reaches the valuation only through the manager’s own process, and that process dampens most of it.

One Australian fund’s board committees had asked why their managers’ discount rates had not risen with government bond yields. In infrastructure, at least, the answer is that valuers often use a long-run risk-free rate rather than the current one. Managers confirmed this to us and defended the smoothed averages as consistent.

A group risk officer described a second mechanism they had tried to stamp out internally. When market conditions would have justified a higher valuation, the extra valuation room was used to recognise asset-specific bad news that had been held back. When conditions worsened, unbooked good news appeared instead. As a result, the headline figure barely moved and because the offset sits in projected cash flows, watching discount rates alone will not reveal it.

Property is slower to react to market movements. One chief investment officer called Australian property valuation backward-looking. It was anchored to past sales and reluctant to move without new ones. In 2020, they noted, funds that wrote property down later had to reverse those write-downs, while funds that waited looked vindicated. As a result, the industry has learned not to move first. Our international survey points the same way with only about a third of practitioners said they revalued assets during periods of market stress.

What makes marks move

Marks move when someone must transact at them, when an observable price contradicts them, or when the question becomes an accounting test with a clearer answer than “what is this worth?”

When investors trade at NAV

When a reported value is not used to price transactions, there is little pressure to keep it current. Once it sets the price at which investors enter or exit, that pressure is immediate. Australian funds that price members daily watch index-based triggers and will adjust a manager’s valuation in their own books before the manager does. Managers whose investors do not transact at NAV rarely revalue between scheduled dates.

There is a complication. One Australian valuations manager explained why their fund would not mark its private equity holdings down to secondary-market discounts: doing so would transfer wealth from existing investors to new ones. Fixing a stale value can create a fairness problem of its own, which helps explain why stale values persist.

Suspending redemptions is the other side of the same coin. A freeze looks like a liquidity measure, but it is also a statement about valuation. It concedes that the reported NAV can no longer be relied on to price transactions between investors.

When recovery assumptions have to change

Direct lending is mostly floating-rate, so these loans are not very sensitive to interest rates. An insurance investor told us managers price off spread curves but move the mark only when the shift is material; small moves leave loans at par. He described a sub-investment-grade loan being downgraded with no change in its price, on the reasoning that it was senior secured and recovery would be full.

When credit deteriorates, then the only thing that matters is expected recovery. If collateral values fall across a concentrated portfolio, recovery assumptions have to change, and those flow into expected credit loss provisions rather than the headline valuation. For an auditor, this is firmer ground. Whether credit deterioration has been recognised as the accounting standards require is far easier to test than what an illiquid loan is worth.

Why Metrics was different

Auditors review valuations routinely. Forcing a material revision is rare. At Metrics, the pressures described above arrived at once.

Metrics’ ASX-listed trusts trade continuously, so investors had a market price to set against the reported NAV. Meanwhile, the underlying wholesale funds stopped publishing NAVs and suspended applications and redemptions until the audit was complete. In effect, that was the admission described above: for a time, the NAV could not be used to price transactions between investors. The regulator was already watching the sector closely. And the review reportedly centred on expected credit loss provisions for distressed assets, where an auditor’s question has the clearest answer.

The valuation techniques at Metrics were much like those used elsewhere. What set this case apart was the incentives: an observable market price, pressure on NAV-based transactions, regulatory attention, and an accounting test the auditor could apply.

What to watch

• Concentration. Around half of Australian private credit is property-backed, a far larger share than in the United States. When collateral values move together, provisioning models calibrated on diversified historical losses are likely to understate expected losses.
• The range model. If a range cannot rule out any estimate inside it, the questions that matter are how wide ranges are, who sets them, and what obliges a manager to justify where the reported figure sits. None of this is publicly disclosed.
• Whether supervision reaches measurement. APRA is asking funds for exposure data, not asking managers to justify their valuation methods. Across the seven regimes we reviewed, supervisors tend to examine governance, controls and processes rather than test valuations. A process can be inspected; whether a judgement-based valuation is correct generally cannot.

We will set out recommendations on these points in a forthcoming paper.

At Metrics, the process was followed and two external firms provided input. It still took an auditor, under unusual pressure, to move the number.