ADI Housing Books Grew $5.6bn in July, the Weakest Month Since February 2024. Business Credit Added $7bn

On 31 August, APRA published its Monthly Authorised Deposit-taking Institution Statistics for July 2026. The broker press covered it as two stories: mortgage growth easing across the top ten ADIs, and CBA closing the gap on NAB in business lending. Open the data file and they are the same story.

In July, the entire ADI system added $5.61 billion to its housing book. In the same month it added $7.01 billion to its loans to non-financial businesses — from a base roughly half the size. Business credit grew faster than housing in dollars and, at 0.56 per cent against 0.22 per cent, by a wide margin in percentage terms.

Business credit outgrowing housing in a single month is not unprecedented. On APRA’s back-series it has happened in nine of the last twenty-six months. What is unusual is the housing figure itself. At $5.61 billion, July was the weakest month of net housing book growth since February 2024.

What the July file actually shows

At 31 July 2026, ADIs held $2,511.4 billion of housing loans on their Australian books — $1,690.0 billion owner-occupied and $821.4 billion investment. Loans to non-financial businesses stood at $1,269.4 billion.

Over the month, owner-occupied lending grew 0.27 per cent and investment lending 0.13 per cent, for a combined housing figure of 0.22 per cent. Business lending grew 0.56 per cent.

Over the twelve months to July, the gap is wider and far more consistent. Housing books grew 6.76 per cent; business books grew 10.03 per cent. In dollars, housing still added more across the year — $159.0 billion against $115.8 billion — because it starts from a base roughly twice the size. Percentage growth is the better read of appetite, and on that measure business credit has been compounding around half again as fast as housing for a full year.

Investment lending is the one part of the housing book still growing at a respectable annual rate: 8.68 per cent, against 5.85 per cent for owner-occupied. But it was also the weakest component in the month, at 0.13 per cent. The annual number describes what was written through late 2025 and early 2026. The monthly number describes what is happening now, and the two are diverging.

Every major, the same pattern

The four majors reported these balances on their Australian books at 31 July 2026, with growth measured against July 2025:

• CBA — housing $637.5 billion, up 6.99 per cent; business $247.7 billion, up 13.92 per cent.
• Westpac — housing $518.0 billion, up 6.06 per cent; business $210.4 billion, up 12.99 per cent.
• NAB — housing $351.5 billion, up 4.70 per cent; business $276.5 billion, up 10.91 per cent.
• ANZ — housing $331.9 billion, up 4.31 per cent; business $157.8 billion, up 5.50 per cent.

Every one of the four grew its business book faster than its housing book over the year. For CBA and Westpac, the business book grew at roughly twice the rate of the mortgage book. NAB, which already holds the largest business book of the four, still grew it more than twice as fast as its mortgages.

The mortgage market is not one market

The $5.61 billion system figure conceals an unusually wide spread. Among ADIs with housing books above $20 billion, twelve-month growth ranged from Macquarie Bank at 25.85 per cent down to Bank of Queensland at minus 8.40 per cent — a spread of more than 34 percentage points.

Between those poles: ING Bank (Australia) grew 10.65 per cent, Heritage and People’s Choice 9.33 per cent and HSBC Bank Australia 4.64 per cent, while Norfina (the entity formerly named Suncorp Bank) was effectively flat at 0.11 per cent, Bendigo and Adelaide Bank at minus 0.02 per cent and AMP Bank at minus 0.13 per cent.

A “flat market” is not flat at the lender level. Several sizeable books are shrinking in absolute terms, which means those lenders are losing more loans to discharge and refinance than they are writing. Others are buying share aggressively. If a large share of your settlements sits with a lender in the second group, the competitive pressure you feel over the next two quarters will be very different from a broker concentrated in the first.

Three things this data cannot tell you

This is a balance-sheet publication, and it is worth being precise about its limits before drawing conclusions from it.

A book balance is a stock, not a flow. Net growth of $5.61 billion is new lending minus repayments, discharges, refinances out and amortisation. A lender writing record volume can still report a flat book if its back book is running off just as fast. Nothing in these figures tells you directly how much new lending was written in July.

Non-bank lenders do not appear. The publication covers authorised deposit-taking institutions. Non-bank lenders are not ADIs, so whatever proportion of broker flow goes to them is entirely absent from these numbers. Any conclusion drawn about “the market” from this file is a conclusion about the ADI portion of it.

“Loans to non-financial businesses” is a very wide category. It spans a $200,000 equipment facility written through a broker and a syndicated corporate facility written in an institutional banking division. Several of the fastest-growing business books in the July file belong to foreign wholesale banks with no retail or broker presence in Australia. The category is not a proxy for broker-addressable SME demand, and should not be read as one.

Why a balance-sheet split matters to a broker

Brokers do not lodge files into a balance sheet. But the place a bank is finding its growth is usually the place its internal resources have already moved to, and that shows up in the things brokers actually experience.

Assessment resourcing follows growth targets. So does BDM coverage, credit escalation appetite, the willingness of a credit officer to look properly at an exception rather than decline it on policy, and the speed at which a file moves from lodgement to conditional. When a division is comfortably ahead of target, it has room to be flexible. When it is behind, the same file can meet a harder answer.

If the July data is telling you anything about the second half of 2026, it is that the mortgage divisions of most ADIs are working a slower-growing book, while their business divisions are running hot. For a residential broker, that suggests sharper competition for a smaller pool of clean deals, and continued pricing aggression on retention. For a broker with genuine commercial or asset finance capability, it suggests the opposite: a set of lenders actively looking for volume.

That is a hypothesis about your panel, not a conclusion about your files. It is also a specific, answerable question to put to a BDM, which is more than most market commentary offers.

The demand picture sitting behind it

Two other recent datasets point the same direction, though each measures something different and they should not be conflated.

On enquiry demand, Equifax figures reported by The Adviser on 11 August recorded four consecutive months of year-on-year decline in mortgage demand — down 0.9 per cent in April, 6.6 per cent in May, 18.8 per cent in June and 16.4 per cent in July, with first home buyer demand down 19.1 per cent in July. Equifax’s Chief Solutions Officer, Kevin James, was reported as describing demand as settling into a lower baseline rather than reacting to individual rate decisions.

On settlement flow, the MFAA reported on 3 September that brokers arranged a record 81.6 per cent of new residential home loans in the June 2026 quarter, on $139.08 billion of new lending, using data from Cotality. That is the channel’s highest recorded share.

Three different lenses — enquiry demand, settlement flow and net book stock — measuring three different things across overlapping periods. They cannot be added together or directly compared. But they are not telling contradictory stories: the broker channel is capturing a record share of residential flow at the point where the underlying pool is growing more slowly than it has in over two years.

Five checks to run before 30 September

APRA publishes the August figures on 30 September. That gives roughly four weeks to do something more useful than read the next set of numbers.

1. Map your settlements against book direction. Rank your last twelve months of settlements by lender, then put each lender’s twelve-month housing book growth beside it. If most of your volume sits with lenders whose books are flat or shrinking, you are exposed to whatever repricing or retention campaign they run next — and you will feel it first.
2. Ask each BDM one precise question. Not “how’s the market”. Ask where their bank’s growth target sits for this half, whether assessment resourcing has moved with it, and what their current lodgement-to-conditional times actually are. The answers tell you more about your next quarter than any published statistic.
3. Test whether your commercial accreditation is real. Accredited is not the same as active. If you hold commercial or asset finance accreditations you have not used in twelve months, you do not have a diversified business — you have a diversified list. Establish whether you can write the deal properly yourself, or whether the right answer is a referral arrangement with someone who can.
4. Look at your own runoff, not just your settlements. The system-level lesson of this data is that a book can be flat while a great deal of lending happens. Run the same calculation on your own trail book. Net growth, not gross settlements, is the number that describes your business.
5. Diarise the next release. 30 September, August data. The question to carry into it is whether July’s $5.61 billion was a single soft month or the start of a trend. One month is not a trend, and this article does not claim it is.

A note on scope

If this data prompts you to look harder at commercial or business lending, it is worth being clear-eyed about the regulatory difference rather than assuming your residential process transfers across. Consumer credit assistance and genuine business-purpose lending sit under different parts of the framework, and the obligations that attach to each are not identical — including the best interests duty, which sits in the consumer credit regime.

This article does not attempt to state what applies to any particular file, and a business purpose declaration is not a substitute for getting the characterisation right. If you are moving into commercial lending, or increasing the share of it in your business, confirm your obligations and your process with your licensee or aggregator compliance team, and take independent advice where the position is unclear.