ADI Housing Credit Fell for the First Time in APRA’s Series. Strip Out One Lender and It Grew $6.3bn

The number that will get quoted, and the number that matters

APRA released its Monthly authorised deposit-taking institution statistics for August 2026 on 30 September. On the face of it, the release contains a genuinely unusual data point: the combined housing book of Australia’s authorised deposit-taking institutions fell.

Owner-occupied and investment lending on the Australian books of the 119 ADIs that reported came to $2.504 trillion in August, down $7.77 billion — or 0.31 per cent — from July. Checked against APRA’s own back-series, which runs from March 2019, that is the only monthly fall in the entire period. Every other month in more than seven years went up.

That is the number that will get quoted. It is also, on its own, close to meaningless for the way you place a file.

Strip out a single institution and the picture inverts. Westpac Banking Corporation’s housing book fell $14.12 billion in the month, from $517.96 billion to $503.84 billion — a 2.73 per cent drop in thirty-one days. Excluding Westpac, the rest of the ADI sector grew its housing book by $6.35 billion, or 0.32 per cent, which is an ordinary month by recent standards.

So the sector did not stop lending in August. One balance sheet moved, and it moved by enough to drag a $2.5 trillion aggregate negative. For brokers, the useful work is in understanding what that one line does and does not tell you — and then in reading the rest of the table properly, because it contains a trap that makes several lenders look far hungrier than they are.

What the Westpac line rules out

A $14 billion single-month move in a mortgage book is not repayments. Australian borrowers do not pay down 2.7 per cent of a major bank’s book in a month. So something structural happened, and APRA’s monthly statistics let you rule out two of the usual explanations.

The first is an on-balance-sheet securitisation shift. It wasn’t that: Westpac’s total securitised assets on balance sheet also fell over the month, from $157.86 billion to $154.33 billion. If loans had been moved into an on-book securitisation vehicle, that figure would have risen.

The second is a transfer to another ADI — the mechanism that explains most large step changes in this series. It wasn’t that either. No other ADI’s housing book grew by anything close to $14 billion in August. The largest single increase in the entire table was Commonwealth Bank at $2.60 billion. Across all 119 reporting institutions, the gains do not add up to a destination.

Westpac’s total resident assets fell $28.23 billion over the same month, so the housing movement sits inside a broader balance-sheet contraction. And APRA’s “Revisions” worksheet for this edition lists no restatement of Westpac’s figures — the numbers are as published, not a correction of earlier data.

Beyond that, the monthly statistics do not explain the cause, and it would be wrong to guess at one. APRA’s own explanatory notes caution that the data in this publication differs from what ADIs disclose in their financial statements and profit announcements, and APRA expressly disclaims responsibility for the accuracy and completeness of the material. A movement of this size could reflect a portfolio sale, an internal transfer, a change in reporting treatment, or a combination — the publication does not say, and at the time of writing no public explanation from Westpac addressing the August figure had been published that this masthead could locate. Treat it as an open question rather than a signal about the bank’s appetite for your next deal.

One point of definition matters here, and it is one brokers routinely miss. Westpac Banking Corporation’s return is the group’s multi-brand Australian book. St.George, BankSA and Bank of Melbourne do not report separately in this publication — they are absent from the list of 119 reporting ADIs — so their lending sits inside that single Westpac line. The figure is not the Westpac-branded book.

The trap in the growth column

The more useful part of the release is the twelve-month view, because that is where credit appetite shows up. Over the year to August 2026, ADI housing credit grew 5.90 per cent, or $139.59 billion. Owner-occupied lending grew 5.15 per cent and investment lending 7.50 per cent.

Rank individual lenders by twelve-month growth, though, and the table appears to say something startling. MyState Bank up 78.37 per cent. Bank of China (Australia) up 27.39 per cent. Macquarie Bank up 24.83 per cent. Teachers Mutual Bank up 24.10 per cent. Bank Australia up 18.09 per cent.

Four of those five did not grow that way by writing loans. They grew by absorbing another ADI’s licence — and the same dataset proves it, because when one ADI’s book is consolidated onto another’s licence, the first one simply stops reporting.

• MyState Bank jumped $4.37 billion in December 2025. Auswide Bank Ltd’s final appearance in the series is November 2025, with a $4.26 billion housing book. MyState and Auswide merged, and the data shows the licence consolidation.
• Teachers Mutual Bank jumped $1.44 billion in May 2026. Australian Mutual Bank Ltd stops reporting that month, with $1.37 billion.
• Bank Australia jumped $4.79 billion in July 2025 as Qudos Mutual Ltd ($4.70 billion) exits the series, and a further $1.58 billion in November 2025 as Australian Unity Bank Limited ($1.41 billion) exits.

None of those lenders did anything improper, and none of this is hidden — but a broker scanning a growth league table and concluding that MyState is writing loans at nearly eighty per cent a year has misread the table by a wide margin. Excluding the merger step, MyState’s book has moved from $10.56 billion to $11.33 billion across eight months: real growth, at a fraction of the headline.

The test is simple and takes a minute. Look at the largest single-month change inside the twelve-month window. If one month accounts for most of the annual movement, you are looking at a corporate transaction, not credit appetite.

Who is actually lending

Apply that test and a much shorter, more useful list survives — lenders whose books rose steadily, month after month, with no step change.

Macquarie Bank is the standout. It added $37.38 billion over the year, taking its book to $187.91 billion — growth of 24.83 per cent that is entirely organic. Its biggest single month in the window was $3.94 billion, which is roughly an average month for it; it added between $1.98 billion and $3.94 billion in every one of the last fourteen months. In dollar terms it added almost as much as Commonwealth Bank ($41.04 billion) off a book less than a third the size.

Among the majors, CBA is the only one growing meaningfully faster than the market: 6.85 per cent against system growth of 5.90 per cent. ANZ grew 4.39 per cent, NAB 4.31 per cent and Westpac 2.82 per cent — all below system, which means all three lost share over the year.

Further down, the consistent organic growers include ING Bank (Australia) at 9.90 per cent, Bank of China (Australia) at 27.39 per cent off a smaller base, Newcastle Greater Mutual Group at 9.85 per cent, Heritage and People’s Choice at 9.12 per cent and Beyond Bank at 12.07 per cent.

The other end of the table is where the broker-relevant risk sits. Three lenders are in sustained run-off:

• Bank of Queensland — down 7.57 per cent, or $4.15 billion, over the year, with a decline in every single month of the period.
• AMP Bank — down 1.69 per cent, after growing through 2025. It has fallen in six of the last seven months, with consecutive falls in every month since April 2026.
• Norfina Limited — down 0.34 per cent, and falling every month since March 2026.

Bendigo and Adelaide Bank, at 0.52 per cent growth, is close to flat against a market growing more than eleven times faster.

Why a run-off book matters more this quarter than last

A shrinking book is not a scandal. Lenders make deliberate decisions to prioritise margin over volume, and a lender in run-off today can be competing hard in six months. But the timing of this release makes those three lines worth a broker’s attention, for a specific reason.

On 29 September the RBA’s Monetary Policy Board increased the cash rate target by 25 basis points to 4.60 per cent — its fourth increase of 2026. The Board said inflation “remains elevated and some of the upside risks flagged in August are materialising”, and that “a further tightening in financial conditions is warranted to support a return of inflation to target”.

Higher assessment rates compress borrowing capacity, and the effect on refinancing is already visible in broker-reported data. The MFAA’s August 2026 Market Sentiment Survey of 588 brokers, as reported by The Adviser, found 49.2 per cent of brokers seeing an increase in clients unable to refinance because of serviceability — up from 24.4 per cent six months earlier. MFAA chief executive Anja Pannek said: “Responsible lending must remain at the centre of our system. At the same time, borrowers who have consistently met their repayments should not be unnecessarily prevented from moving to a more affordable or suitable home loan.”

Put the two datasets together and the exposure is obvious. A client sitting in a book that is running off is in a portfolio with no commercial reason to compete for their business — and, if their serviceability has deteriorated, limited ability to leave. That is the combination that produces a trapped borrower: a lender with no incentive to sharpen the rate, and a borrower who cannot pass another lender’s assessment to escape it.

You can identify those clients today, from your own CRM, before they call you.

What to review this week

A practical sequence, in roughly the order it is worth doing:

1. Segment your back book by lender. Pull every active loan in your book and group it by funder. You are looking for concentrations with the three run-off lenders above, and with any lender growing well below the 5.90 per cent system rate.
2. Flag the clients who are both exposed and constrained. Within those concentrations, identify the loans written in the last two to three years at higher LVRs. Aussie Home Loans data, as reported by The Adviser, found 20.7 per cent of borrowers who purchased between July 2023 and August 2025 had an LVR above 80 per cent; Aussie chief executive Sebastian Watkins said that “once you push above 80 per cent, the door to a competitive refinance can start to slam shut”. Those are the files where a repricing conversation matters more than a refinance pitch, because a refinance may not be available.
3. Run the retention play before the rate review. For clients you cannot move, a pricing request with the existing lender is the realistic lever. The MFAA survey found 96 per cent of brokers had helped clients obtain discounts from their lender in the previous six months — it is standard practice, and it is more productive than a refinance application that will decline.
4. Re-test your placement assumptions. If your panel habits were formed when the majors were growing at system or better, note that three of the four grew below system over the past year. That is not a reason to move a file — but it is a reason to check current policy and pricing rather than rely on where a lender sat eighteen months ago.
5. Document the reasoning, not the conclusion. Whichever way a file goes, the file note needs to show why that lender suited that client’s circumstances and objectives.

Reading the APRA release yourself, in ten minutes

The publication is free, and the back-series file is more useful than the monthly one. Download the back-series workbook from APRA’s website, open the single data table, and filter to the lenders on your panel. Add owner-occupied and investment housing together for a total book. Then compare the latest month to the same month a year earlier for the annual change, and scan every intervening month for a single large jump or drop. If you find one, check whether an ADI stopped reporting in that month — that is almost always your explanation.

The diversification number hiding in the same release

One more figure is worth noting, because it sits in the same table and points in a different direction. Over the year to August 2026, ADI lending to non-financial businesses grew 10.30 per cent, reaching $1.279 trillion. Housing grew 5.90 per cent.

Business credit is expanding at nearly twice the rate of housing credit on the same balance sheets. That is not a reason to pivot a residential business overnight, and commercial lending carries different accreditation, skill and risk requirements. But if you have been weighing diversification as a response to compressed residential volumes, the lenders’ own books are telling you where the growth currently is.

What to watch next

September’s edition of the monthly statistics, due at the end of October, will show whether the Westpac movement was a one-off adjustment or the start of a trend — and whether the ADI aggregate returns to growth, which on the ex-Westpac arithmetic it should. Westpac’s financial year ended on 30 September, so its full-year disclosures are the more likely place for an explanation of the August figure than APRA’s statistics, which do not carry commentary.

Also worth watching: whether the sub-system growth at ANZ, NAB and Westpac prompts sharper pricing or policy from any of them, and whether the run-off lenders re-enter the market competitively, which would change the retention calculus for clients currently sitting in those books.

The takeaway

A headline fall in national housing credit turned out to be one institution’s balance sheet, and a league table of the fastest-growing lenders turned out to be mostly mergers. Both are reminders that aggregate data is where misreadings get made, and that the lender-level detail — which APRA publishes free, every month — is where the information a broker can actually use sits.

The two lists worth keeping are short. The lenders compounding organically, led by Macquarie and CBA, are the ones buying new business. And the lenders in sustained run-off — Bank of Queensland, AMP Bank and Norfina — are where your retention risk is concentrated, particularly for clients whose serviceability has moved against them since the cash rate reached 4.60 per cent. Those clients are in your CRM right now. The question is whether you find them before their next rate notice does.