Same-Lender Refinancing Fell 28% — More Than Three Times the 8.8% Fall in Switching. The Steepest Drop in the Release Was the Easier Option

The two numbers

Two figures in Equifax’s September data describe the same refinance market and move at very different speeds. Refinancing with a different lender fell 8.8% year-on-year. Refinancing with the same lender fell 28%.

Both are declines, so it would be easy to file them together as more evidence of a slow market. That would waste them. The second number is more than three times the first, and on the reporting of the release it is the steepest fall of any segment Equifax measured. The gap between those two numbers is more useful to a broker than either number alone, because it says something specific about where the remaining work in an existing client book actually sits.

It also runs against the intuition. Refinancing with the existing lender is normally the lower-friction path: no discharge, no new lender’s credit assessment from scratch, often no revaluation. When conditions tighten, the easier path is the one you would expect to hold up better. In September it did the opposite.

What else the September release says

The figures come from Equifax’s Consumer Market Pulse for the September 2026 reference period, reported across the trade press in the week to 15 October. Overall mortgage demand fell 19.2% year-on-year — a sixth consecutive monthly decline, and on MPA’s and Australian Broker’s reporting the weakest reading since October 2022. August had been down 14.1%.

The state picture is uniformly negative on overall demand: NSW down 21.2%, the ACT down 21%, Queensland down 19.2%, Victoria and South Australia down 17.9% each and Western Australia down 16.6%. Within the switching figure, MPA reported Tasmania as the only state to grow, up 1.5%, with NSW the weakest. Within the same-lender figure, NSW fell 31.3%.

First home buyers recorded the largest annual decline in that segment since December 2022, at 22.1%, while the average first home buyer loan sat at a record $743,000 — $634,000 for 18 to 25 year olds and $736,000 for 26 to 35 year olds. Investor demand, measured as borrowers with two or more mortgages, fell 18% after falls of 11.4% in August and 15.1% in July.

Why the gap is the interesting part

Equifax’s own explanation for soft refinancing is serviceability. On MPA’s reporting, James said that in the current climate strict regulatory serviceability assessments appear to be holding borrowers in place; The Adviser reported him describing those assessments as effectively locking borrowers into their current arrangements.

That explanation does real work for the external number. A borrower who cannot clear a new lender’s assessment does not switch, and the 8.8% fall in switching is consistent with that.

But follow the logic one step further and a question opens up. If borrowers are being held in place with their current lender, the same-lender channel is where you would expect that displaced activity to surface — as repricing requests, product switches and restructures. Instead that channel fell 28%, and it has been the weaker of the two for at least two consecutive readings. Banking Day reported the year-to-August figures as same-lender down 23% against external down just 0.8%.

Equifax has not published an explanation for that second part, and nor has anyone else, so what follows is interpretation rather than reported fact. Three possibilities are worth a broker’s attention, and they point to different responses.

The measurement explanation. Demand indices of this kind are built on credit application and enquiry activity. Where a lender reprices a loan proactively through a retention desk, there may be no new application recorded at all. A fall in measured same-lender demand is therefore not necessarily a fall in same-lender repricing. If lenders are being more forward about discounting to hold a book in a shrinking market, measured internal demand would drop even as actual repricing held up.

The assessment explanation. Not every same-lender change is a light-touch exercise. APRA’s prudential practice guide for residential mortgage lending expects a new serviceability assessment whenever there are material changes to loan conditions. Where that applies, the easy internal path is not easy at all — it is the same test the borrower would face elsewhere, which would push internal and external activity down together.

The base effect. Year-on-year comparisons are only as flattering as the year they are measured against. A period of heavy repricing activity in the comparison window would depress the current reading without anything new happening at all.

A broker cannot resolve which of these dominates from published data. What a broker can do is notice that all three point the same way operationally: the client is less likely to start the conversation than they were a year ago.

What APRA’s framework actually says

Because serviceability is doing so much of the explanatory work here, it is worth being precise about what the framework actually requires, rather than what it is often described as requiring.

APRA’s Prudential Practice Guide APG 223 Residential Mortgage Lending, in the version currently published and dated 19 June 2025, states that authorised deposit-taking institutions must apply a buffer over a loan’s interest rate of “at least 3.0 per cent, unless determined otherwise by APRA”. The buffer applies to the rate the borrower will actually pay and excludes introductory or honeymoon rates.

The same guide addresses changed loan conditions directly. APRA expects an ADI to undertake a new serviceability assessment “whenever there are material changes” to loan conditions. The examples it gives include switching from principal and interest to interest-only, extending an interest-only period, switching between fixed and floating rates, and extending the loan term. A new assessment is expected even where immediate repayments fall, if total lifetime repayments increase. That is the mechanism behind the assessment explanation above, and it is why a same-lender restructure is not automatically the soft option a client may assume it is.

APG 223 also defines the exception route. An override, in the guide’s words, “occurs when a residential mortgage loan is approved outside an ADI’s loan serviceability criteria”. Overrides are expected to be captured, subject to risk limits and a defined approval process, and reported regularly.

APRA set out its expectations on the scale of that route in a letter to all ADIs dated 9 June 2023, signed by Chair John Lonsdale. The letter states that banks can use exceptions to policy “if these are managed prudently and limited”, notes that serviceability exceptions have historically made up roughly 2 to 3 per cent of banks’ total housing lending, and says that “prudent banks would have acceptable reasons and clear justifications for loans written outside policy”. Banks reporting large volumes, it says, will be subject to heightened supervisory attention.

The practical point for a broker is narrower than it first appears, and more useful. A client who “cannot refinance” has usually been told that by one lender, on one policy, at one point in time. The buffer is a floor on the assessment, not a uniform outcome across a panel, and the assessment can turn on things that change — a completed probation period, a cleared liability, a corrected living expense figure, a different lender’s treatment of the same income.

Where the remaining capacity sits

The age breakdown in the September data is the clearest practical signal in the release. On the figures reported by Australian Broker and The Adviser, the cohorts reported for lender switching were these three: up 3.2% among 56 to 60 year olds, up 2.3% among the over-60s, and down 14.2% among 26 to 35 year olds. Figures for the remaining cohorts were not reported.

No one has published the reason, but the profile is suggestive. Borrowers in their late fifties and sixties are more likely to hold lower loan-to-value ratios, longer-held equity and smaller loan balances relative to income — the characteristics most likely to clear a full assessment with a 3 per cent buffer applied.

That is an opportunity, with a caveat attached. Older borrowers bring considerations that younger ones do not: loan terms extending towards or past retirement, exit strategy, and the lender’s own approach to assessing both. Those are matters lenders assess specifically and that a broker has to document. The one cohort the data shows growing is also the cohort that carries the most evidencing with it. It is a genuine segment, not a shortcut.

What this changes on your desk

A 28% fall in same-lender activity, whatever its cause, has one unambiguous implication: materially fewer clients are initiating the conversation than a year ago. Whether that is because they have been declined, because they assume they will be, because their lender quietly repriced them, or because they have simply stopped thinking about it, the practical effect on a brokerage is the same. The flow of inbound review requests that a book generates on its own has thinned.

That makes the scheduled annual review the mechanism that surfaces the difference, rather than a courtesy call. It is where a broker establishes which clients genuinely cannot move, which could move but have not asked, and which have already been repriced without being told what else was available.

It also matters that the recommendation to stay put is still a recommendation. A file that shows a client was reviewed, options were considered at their current borrowing capacity, and staying was the better outcome is a different file from one with a year of silence on it.

A four-step review of the retention book

What to watch next

• The October reading. The September data largely predates the full effect of the cash rate increase to 4.60 per cent on 29 September 2026. Whether the same-lender channel keeps falling once that decision works through enquiry is the next useful datapoint.
• Whether the over-55 growth holds. Growth of 3.2% and 2.3% is a thin margin. If it persists it is a segment worth building around; if it reverses it was a timing artefact.
• The gap itself. If same-lender and external refinancing converge, the measurement explanation looks weaker. If the gap widens, proactive lender retention becomes the more likely driver — and that is a direct competitive issue for the broker channel.
• Any movement on the buffer. The 3 per cent minimum is expressed in APG 223 as applying unless APRA determines otherwise, which is the mechanism by which it would change.

The bottom line

The headline from Equifax’s September release is a 19.2% fall in mortgage demand, and that number will be quoted widely. The more useful number for a broker is the 28% fall in same-lender refinancing, because it is both the steepest in the release and the one that contradicts what you would expect in a market where borrowers are reportedly stuck.

Whatever is driving it, the operational consequence is settled: clients are not starting the conversation. In a shrinking market the brokerages that hold their volume will be the ones that already know which clients in their book can move, which cannot, and what would have to change for the second group to join the first — and that have the file to show for it.