The Refinance Split Nobody Put Side by Side
Equifax Consumer Credit Demand Market Pulse — mortgage demand, year to August 2026
Two numbers from the same month
Refinancing enquiry volumes, August 2026 against August 2025.
Refinancing with the existing lender
Banking Day reported the same series at −23%
Refinancing with a different lender
Banking Day reported −0.8%
The gap opened in three months
Year-on-year change in refinancing demand, same lender against different lender. Bar length shows the size of the fall.
The age split
Year-on-year change by age, year to August 2026. The two outlets reporting this release used different age brackets, so these figures sit alongside each other rather than forming one series.
What the official series confirms
ABS Lending indicators, June quarter 2026, released 14 August 2026. The ABS measures external refinancing — switches to a different lender — so internal refinancing does not appear in it at all.
The broker takeaway
Borrowers have not stopped refinancing nearly as fast as they have stopped refinancing where they already are. Whatever sits behind that, the activity that held up in August is the activity that runs through a broker — and the only age bands showing growth are the oldest ones, which changes both the conversation and the file.
Sources: Equifax Consumer Credit Demand Market Pulse (year to August 2026) as reported by Australian Broker, 25 September 2026, and Banking Day, 16 September 2026; the May 2026 edition as reported by The Adviser, 16 June 2026; Australian Bureau of Statistics, Lending indicators, June quarter 2026.
Refinancing With the Same Lender Fell 22.8%. Switching Lenders Fell 1.1%. In May the Gap Between Them Was Under Three Points
Equifax’s latest monthly read on mortgage demand contains two numbers that belong next to each other and have mostly been reported apart. In the year to August 2026, refinancing enquiries with a borrower’s existing lender fell 22.8%. Refinancing enquiries with a different lender fell 1.1%. Three months earlier those two series were less than three percentage points apart.
What’s in this piece
What Equifax actually reported
Overall mortgage demand in Australia fell 14.1% year on year in August 2026, the fifth consecutive month of contraction. It was, on the face of it, a better month than July, which recorded a 16.4% fall. New South Wales had the steepest decline at 15.9%; Western Australia the shallowest at 10.4%. First home buyer applications fell 20%. Total refinancing demand fell 12%.
Underneath that refinancing figure sits the split worth your attention. Refinancing with the existing lender fell 22.8%. Refinancing with a different lender fell 1.1%.
A note on sourcing before we go further. Equifax’s monthly Consumer Credit Demand Market Pulse is not issued as a public media release, so the figures here come from the trade coverage of it rather than from a primary document. Australian Broker, reporting on 25 September, put the two refinancing series at −22.8% and −1.1%. Banking Day, reporting the same release on 16 September, put them at −23% and −0.8%. The headline demand figure of −14.1% is identical in both. The gap between internal and external refinancing is therefore somewhere around 22 percentage points rather than exactly 21.7, and nothing in this article turns on the difference.
One more distinction that matters more than it sounds. These are demand figures — mortgage enquiries and applications recorded against consumer credit files. They are not settlements. A borrower who never lodges an application never appears, and a fall in enquiries tells you about intent and activity at the front of the funnel, not about funded volume three months later.
The gap opened in three months
This is not a long-running structural feature of the market that someone has just noticed. It is new, and you can date it.
The May 2026 edition of the same Equifax series, reported by The Adviser in mid-June, showed refinancing with the same lender down 7.0% year on year and switches to a different lender down 4.2%. Both negative, both modest, less than three points between them. Go back to January and the picture inverts: internal refinancing demand was up 16.2% year on year while external switching was up 8.6%. In other words, at the start of 2026 borrowers were going back to their own lender faster than they were going to market.
| Refinancing demand, year on year | Same lender | Different lender | Gap |
|---|---|---|---|
| January 2026 | +16.2% | +8.6% | 7.6 pts |
| May 2026 | −7.0% | −4.2% | 2.8 pts |
| August 2026 | −22.8% | −1.1% | 21.7 pts |
Across seven months the internal series moved roughly 39 points, from strong growth to a heavy contraction. The external series moved about ten points and finished close to flat. Whatever has happened to Australian mortgage demand this year has happened overwhelmingly to the half of it that never leaves the incumbent lender.
Equifax’s own reading of the cycle is that it may be bottoming. Moses Samaha, Equifax Australia’s executive general manager, said the market “may well be at or near rock bottom in terms of this contraction cycle”, while cautioning that further rate increases could push that threshold lower. On first home buyers, he was blunter: “policy incentives alone aren’t moving the needle.”
“We may well be at or near rock bottom in terms of this contraction cycle.”
Moses Samaha, Executive General Manager, Equifax Australia, as reported by Australian Broker, 25 September 2026
Four candidate explanations — and why it matters which is right
Equifax’s published summary does not explain the divergence, and it would be easy to write the flattering version: borrowers have given up on their own banks and gone looking for a broker. That may be part of it. It is not the only reading, and the difference between the readings changes how much of your pipeline planning you should hang on this number.
1. The activity moved out of the data, not out of the market
When a lender holds a customer by discounting the existing loan, no new application is lodged and no credit enquiry is generated. Retention handled as a repricing is invisible to bureau enquiry data. If this is the dominant explanation, internal refinancing has not fallen at all — it has simply stopped being measurable, and the pool of genuinely up-for-grabs borrowers is smaller than the headline suggests.
2. Borrowers have stopped asking their own lender first
The sequence many borrowers used to follow — call the bank, ask for a better rate, go to market only if refused — produces an internal enquiry before an external one. If borrowers increasingly skip step one because they no longer expect it to work, the enquiry that used to be internal becomes external. This reading supports the flattering version.
3. The top-up has gone, and the top-up was usually internal
Internal refinances disproportionately involve an increase — equity release, debt consolidation, a renovation. Those applications depend on both servicing capacity and valuation headroom. A dollar-for-dollar switch to a cheaper lender needs far less of either. Equifax’s own quarterly analysis for the June quarter recorded the national average mortgage size contracting by about $8,000, or 1.1%, between March and June, with Brisbane down $15,000, Sydney down $12,000 and Melbourne down $11,000. Smaller loans are consistent with fewer increases.
4. Composition, not behaviour
If the borrowers most likely to refinance internally are also the borrowers whose capacity has tightened most, the mix shifts without anyone changing their mind about anything.
Why this matters practically: explanations 2 and 3 describe demand that a broker can win. Explanation 1 describes demand that has quietly moved behind a lender’s retention desk where no broker will ever see it. The published data does not separate them. Plan on the basis that the opportunity is real but smaller than 22 percentage points implies, and test it on your own book before you spend money on a campaign.
What the ABS can and cannot see
There is a reason this split has not been part of the industry conversation: the official statistics cannot show it.
The ABS Lending indicators series, in its June quarter 2026 release published on 14 August, reports refinancing as external refinancing — refinances to a different lender. Internal refinancing is not in the series. The collapse Equifax is picking up is, by construction, absent from the national accounts of Australian mortgage lending.
What the ABS does show corroborates the other half of the Equifax finding rather neatly. Owner-occupier external refinancing ran at $41.9 billion in the June quarter, down 2.8% on the quarter but up 2.8% on the year. Investor external refinancing ran at $25.2 billion, down 0.8% on the quarter and up 12.0% on the year. Against a backdrop where total new dwelling commitments fell 5.2% over the quarter to $97.6 billion, and investor commitments fell 10.2%, external refinancing was the steadiest large flow in the release.
Two independent datasets, built on different foundations, agree on the external side: switching lenders held up while nearly everything else softened. That is about as much confirmation as a broker gets before a quarter starts.
The age bands that grew
The age breakdown is the second half of the story, and it points in a single direction.
Australian Broker reported demand down 21.7% among 18 to 25 year olds and down 18.1% among 26 to 35 year olds, against a fall of just 1.7% for those aged 66 and over. Banking Day, reporting the same release, recorded demand up 13.3% among 56 to 60 year olds and up 9.2% among those over 60.
Handle these bands carefully. The two outlets used different age brackets, and a rise of 9.2% for the over-60s alongside a fall of 1.7% for the over-66s can only both be true if the 60 to 65 range is strongly positive. The published summaries do not let us reconcile them precisely. What they support is a narrower claim: the growth sits in the pre-retirement band, not among the oldest borrowers.
The swing is recent here too. In the May edition, the 56 to 65 cohort was down 13.5% year on year — the weakest of the five bands reported that month. Three months later the pre-retirement band is the one growing. Notably, that May report also flagged the over-65s recording a 3.3% rise specifically in refinancing with a different lender, which is the earliest sign in this data of the pattern that has since become the headline.
Meanwhile the first home buyers who do transact are borrowing more, not less: Banking Day put the average first home buyer mortgage at $740,000, at all-time highs, which Equifax links to demand for the 5% deposit guarantee. Fewer first home buyers, each with a larger loan.
What this does to your next quarter
Put the two findings together and the shape of available volume in late 2026 is unusually specific: refinancing that leaves the incumbent, skewed towards borrowers in their late fifties and early sixties.
Three consequences follow.
Your retention exposure and your acquisition opportunity are the same number read from opposite ends. If lenders across the market are holding customers by repricing rather than by re-documenting, that is happening inside your trail book too. A loan quietly repriced by the lender keeps paying you trail and generates no upfront — and a borrower who has already had one unprompted conversation with their lender about rate is a borrower whose next review may not involve you at all. An annual review call that lands before the lender’s retention team does is worth more this quarter than it was in January.
The prospecting profile has moved about twenty years up the age range. A 58-year-old refinancer is not a rate-only conversation. The questions that come with that file — remaining term against expected working life, whether to preserve the existing term or reset it, debt consolidation, whether an offset structure beats an extra repayment — are advice-shaped, which is a better fit for a broker than for a call centre, and they carry specific documentation consequences covered below.
First home buyer volume is not coming to the rescue. Down 20% year on year, with Equifax’s own executive saying incentives are not moving the needle, and the loans that do complete are larger. If your model depends on first home buyer flow, the pipeline arithmetic for the next two quarters needs redoing rather than hoping.
The file discipline on a pre-retirement refinance
The following is general information about published regulatory guidance, not compliance advice. Your obligations depend on your licence arrangements, and your licensee or compliance adviser is the right person to apply them to your process.
Two paragraphs of existing ASIC guidance do real work on a file of this shape, and both are worth reading in the original rather than in summary.
On borrowers approaching retirement, ASIC’s Regulatory Guide 209, Credit licensing: Responsible lending conduct, says at RG 209.64(a) that “if a consumer is approaching retirement, and will still be making repayments on the credit product after their expected retirement age, you will need to determine whether this event is likely to change their income”. It sits within the guide’s broader treatment of reasonably foreseeable changes in a consumer’s circumstances. The practical reading is that where the proposed term runs past the borrower’s expected retirement, the income position after that point is something to establish and record on the file, not something to assume.
On refinancing generally, ASIC’s Regulatory Guide 273, Mortgage brokers: Best interests duty, notes at RG 273.58 that “for consumers who are thinking about refinancing an existing loan, the expenses incurred when refinancing may exceed the cost savings of a new loan”. That matters most in exactly the scenario this data points at. Resetting a 30-year term over a loan with 22 years left will lower the monthly repayment and can raise the total cost of the credit substantially — a trade-off that is defensible when the borrower has chosen it with the numbers in front of them, and considerably harder to defend when the file records only the lower repayment.
The honest version of the compliance point: nothing about this data creates a new obligation. What it changes is how often you will be dealing with the files where these two paragraphs already applied. A market where the growth segment is 56 to 65 and the product is a refinance is a market where term, retirement income and cost-over-the-life-of-the-loan turn up on a much larger share of your files than they did last year.
What to review this week
- Pull the age profile of your own book. How many clients are between 54 and 64? That is the cohort the market data says is active. Most brokers have never segmented on age and have no idea what the answer is.
- Test explanation one against your own numbers. Of the loans in your book repriced in the last six months, how many were repriced by the lender without you raising it? That ratio tells you, better than any national figure, how much retention activity is happening around you.
- Check your review cadence against your largest trail positions. If a lender’s retention team gets to a client before your annual review does, the conversation starts from their number.
- Look at your last ten refinances where the term was reset. Does each file show a total-cost comparison and the borrower’s reasoning, or only the new repayment? RG 273.58 is the relevant paragraph.
- Add a retirement-income prompt to your fact find where the term crosses expected retirement. Not a tick box — a recorded answer. RG 209.64(a) is the relevant paragraph.
- Reforecast first home buyer volume down, not flat. Five consecutive months of contraction and a 20% annual fall in applications is a trend, not a wobble.
Key takeaways
- In the year to August 2026, Equifax recorded refinancing enquiries with the existing lender down 22.8% and refinancing with a different lender down 1.1%. Banking Day reported the same series at −23% and −0.8%.
- That gap is new. In May the two series were 2.8 points apart; in January, internal refinancing was growing faster than external.
- These are enquiry and application volumes, not settlements, and the published data does not explain the divergence. Lender repricing that generates no new application would produce the same pattern without any borrower changing behaviour.
- The ABS measures external refinancing only, so it cannot show the internal fall — but its June quarter figures ($41.9bn owner-occupier, up 2.8% on the year; $25.2bn investor, up 12.0%) independently support the finding that external switching held up.
- Demand growth was reported only in the older bands — up 13.3% for 56 to 60 year olds and up 9.2% for the over-60s — against falls of 21.7% for 18 to 25 year olds and 18.1% for 26 to 35 year olds. The outlets used different brackets, so the precise shape is uncertain.
- Practically: review cadence on your own book matters more this quarter, the active prospect is older than your average client, and RG 209.64(a) and RG 273.58 will apply to a larger share of your files.
Common questions
Does a 22.8% fall in internal refinancing mean those borrowers are now available to brokers?
Not necessarily, and this is the most important caveat in the data. Bureau enquiry data only records activity that generates a credit enquiry. If a lender retains a customer by discounting the existing loan without a new application, that retention never appears in the series — so a falling internal figure is consistent with retention activity increasing, not decreasing. Some of the fall is probably genuine displaced demand and some of it is probably measurement. Equifax’s published summary does not separate the two.
Why doesn’t the ABS data show this?
Because the ABS Lending indicators series reports external refinancing — refinances to a different lender — and internal refinancing is not included in it. The internal series simply has no official counterpart. That is why bureau data is the only public window onto this split, and why it is worth reading the monthly release rather than waiting for the quarterly official numbers.
Is the 14.1% fall in overall demand as bad as it sounds?
It was the fifth consecutive month of contraction, but it was an improvement on July’s 16.4% fall, and Equifax’s executive general manager described the market as possibly “at or near rock bottom in terms of this contraction cycle” — with the caveat that further rate increases could lower that floor. It is also a demand figure. Settlements lag enquiries, so the effect on funded volume shows up later than the effect on your enquiry count.
What actually changes on a file for a borrower in their late fifties?
Two things come into focus. First, where the proposed term extends past expected retirement, RG 209.64(a) points to determining whether that event is likely to change the borrower’s income — something to establish and record rather than assume. Second, where the term is being reset, RG 273.58’s observation that refinancing expenses may exceed the savings applies with particular force, because a lower repayment achieved by lengthening the term can increase the total cost of the credit. Neither is a new rule. Both simply apply more often when the active segment is older. How they apply to your process is a question for your licensee or compliance adviser.
Should I change my first home buyer strategy on the back of this?
The data supports reforecasting volume rather than abandoning the segment. Applications were down 20% year on year, and Equifax’s view is that “policy incentives alone aren’t moving the needle”. At the same time, the average first home buyer loan reached a reported $740,000, at all-time highs, which Equifax attributes to demand for the 5% deposit guarantee. Fewer buyers, larger loans — so revenue per settlement is holding up better than settlement count.
Breaking news for modern brokers
Market data, lender policy and compliance developments, read for what they change on your files.
Sources. Equifax Consumer Credit Demand Market Pulse, mortgage demand for the year to August 2026, as reported by Mina Martin, Australian Broker, 25 September 2026, and by Banking Day, 16 September 2026. The May 2026 edition of the same series as reported by The Adviser, 16 June 2026. Equifax Consumer Credit Analysis — Q2 2026, published by Equifax Australia, 4 August 2026, for average mortgage size figures. Australian Bureau of Statistics, Lending indicators, June quarter 2026, released 14 August 2026, for new loan commitments and external refinancing values. ASIC Regulatory Guide 209, Credit licensing: Responsible lending conduct, at RG 209.64(a). ASIC Regulatory Guide 273, Mortgage brokers: Best interests duty, at RG 273.58. Equifax’s monthly Market Pulse is not published as a public media release; monthly figures are attributed to the outlets reporting them, and where those outlets differ the article says so.
Refinance Conversation Router
Four questions about the file in front of you. It returns where that borrower sits in the August 2026 demand data, the conversation worth leading with, and the points the two relevant ASIC paragraphs put on the file.
These are prompts, not advice. The regulatory references are to published ASIC guidance and are included so you can read the paragraph yourself. Whether and how they apply to a given file depends on your licence arrangements and your own process — that is a conversation for your licensee or compliance adviser, not a tool on a website. Market figures are from the Equifax Consumer Credit Demand Market Pulse for the year to August 2026 as reported by Australian Broker and Banking Day, and are demand and enquiry volumes rather than settlements.
Disclaimer: This article is for general information and professional development purposes only. It does not constitute legal, compliance, or financial advice. Brokers should consult their aggregator’s compliance team and, where required, seek independent legal advice regarding their obligations under the National Consumer Credit Protection Act 2009 and ASIC’s responsible lending guidelines.
