NAB’s Own Channel Wrote 50% of Its FY26 Drawdowns, Up From 41%. ANZ’s Broker Share Rose to 71% — the Majors Are Now Pulling in Opposite Directions

Three of the four major banks reported full-year results inside eight days, and in every one of them the mortgage distribution line moved. NAB’s own channel wrote half of its Australian home loan drawdowns in FY26. Westpac’s first-party share of new mortgage flow reached 37.2% in the second half. ANZ committed to up to 50% more lenders in its branches over the next five years — while brokers’ share of its new home lending rose to 71%.

Read individually, each of those is a lender story. Read together, they are a distribution story, and it is not the one the headlines imply. The majors are not walking away from the broker channel. They are arriving at different answers to the same question: which loans do we want, and through whom do we want to buy them. For brokers, the consequence is not a commission scare. It is that lender appetite is becoming segment-specific — and an advertised rate is increasingly a rate for a particular shape of file.

What the three results actually say

NAB (results released 5 November). NAB’s own full-year results summary states that proprietary channel drawdowns “improved to 50% in FY26 from 41% in FY25” for Australian home lending, excluding Ubank. That is a nine-percentage-point swing in a single year, and it puts the broker channel at half of NAB’s settlements, down from 59%, as Australian Broker reported from the same result. The same document describes the bank’s operating expense growth as including “investment in proprietary lenders and frontline bankers to support growth”, and lists “ongoing proprietary channel performance improvement and a more targeted approach in the broker market”. Australian home lending balances grew 3.8% over the year. The Adviser reported the half-year detail: proprietary at 52.5% of new flow in the second half against 47.7% in the prior half, and new originations of $39 billion against $47 billion a year earlier.

Westpac (results released early November). Westpac’s own full-year presentation and investor discussion pack prints first-party flow for the Group at 32.6% in 2H25, 34.0% in 1H26 and 37.2% in 2H26. For the Westpac brand alone the series runs 34.8%, 38.2% and 42.4%. Notably, the portfolio tells a slower story than the flow: proprietary balances were 41.6% of the Australian mortgage book at September 2026, down from 42.7% a year earlier, because a book that size takes years to re-weight. Total Australian mortgage balances rose about 7% over the year, from $497.0 billion to $531.9 billion, with third-party-originated balances growing from $212 billion to $221 billion. The pack’s proprietary focus list is operational rather than promotional: increase brand consideration, optimise lead quality, streamline application requirements, reduce products.

ANZ (results released 9 November). ANZ reported FY26 cash profit of $7,591 million, up 31%. Its investor pack lists “Strengthen proprietary origination” among Australia Retail priorities, alongside “Increase lenders in branches, up to 50% more over next 5 years” and “Invest and train mortgage sales force”. Australia Retail home lending grew 4%, from $264 billion to $275 billion. And yet ANZ is the bank whose broker reliance increased: brokers wrote 71% of new home loans in FY26, up from 68%, with proprietary falling from 32% to 29%, and broker-originated loans now 63% of the portfolio against 61%, according to ANZ’s result as reported by both The Adviser and MPA. The Adviser quoted chief executive Nuno Matos saying that “without compromising our great broker relationships, we will materially invest in and train our own mortgage sales force”.

The counterweight. Macquarie, reporting its half-year to 30 September, took its home loan portfolio to $198.2 billion, up 9%, with more than 95% of originations written through brokers and market share at 7.6% in September against 7.1% in March. Wendy Brown, Macquarie Bank’s head of home loan and commercial broker sales, told The Adviser that “our growth to 7.6 per cent of the home loan market forms part of a longer-term trend”. Against the backdrop of brokers facilitating a record 81.6% of all new residential home loans in the June 2026 quarter — the MFAA’s Cotality-sourced figure — it is the banks with the smallest branch networks that are growing fastest through the channel.

This is re-segmentation, not retreat

It is tempting to read a rising proprietary number as a bank reducing its commitment to brokers. The FY26 set does not support that reading as a channel-wide trend. ANZ’s proprietary investment is happening at the same time as its broker share climbs, because its branch network is being rebuilt from a low base and its mortgage book is still growing. Westpac’s first-party flow is rising while its third-party balances also rise in dollar terms. NAB is the one bank where the shift is unambiguous, and even there the bank is explicit that it still wants broker flow — just not all of it.

What is actually happening is a sorting exercise. Each major is deciding which segments it can win profitably through which channel, and the broker channel is being asked to supply a narrower slice of a bank’s book than it did when every lender wanted volume at any shape.

NAB has published its filter

Most banks leave their appetite implicit and let brokers discover it through declines and repricing. On NAB’s FY26 earnings call, the filter was described out loud. Australian Broker reported chief executive Andrew Irvine saying there “still is good flow in the broker market that we’re choosing to compete for, generally in lower [loan-to-value] bands” and at larger ticket sizes. On brokers themselves, Irvine said: “We’ve identified a number of brokers [who] are providing us with more of the flow that we would want to book,” adding that NAB “basically will become, hopefully, the preferred bank for that subset of brokers”. The Adviser, reporting the same media call, put that subset at roughly 5 to 10 per cent of brokers. (The bracketed words are the outlets’ clarifications, not NAB’s.)

Take that at face value and it is useful information rather than a threat. It tells you the two variables that move your success rate with that lender — LVR band and loan size — and it tells you that the bank’s service proposition is being built around a cohort rather than the whole channel. A broker writing $1.2 million refinances at 60% LVR is in a different relationship with that lender than a broker writing 90% LVR first-home-buyer purchases, even if both sit on the same aggregator panel and see the same rate sheet.

The quiet risk sits in your pricing conversation

The commercial risk in a segmenting panel is not that your clients are declined. It is that you quote a lender’s advertised position to a client whose file does not qualify for it, and then have to walk the number back after the client has anchored on it.

When appetite was broad, the advertised rate was close to the rate your client got. When appetite is defined by segment, the headline number increasingly belongs to a specific LVR band, loan size, security type or borrower profile. Pricing discretion, cashback availability and escalation support can all follow the same segmentation logic. None of that is visible on a comparison table.

The practical fix is unglamorous: separate the lenders you are confident about for a given shape of deal from the lenders you are confident about generally. Most brokers already carry this knowledge informally. Writing it down turns it into something your support staff and newer brokers can use, and something you can evidence.

Where ASIC’s October review intersects

This is also where the distribution story meets a compliance one that landed four weeks earlier. On 13 October, ASIC published REP 840, a review of best interests duty and complaints handling practices across 13 credit licensees forming eight large aggregator groups — groups ASIC said are responsible for about two-thirds of mortgage broker representatives in Australia, using data from 1 April 2023 to 31 March 2025.

In ASIC’s accompanying media release, the regulator said seven of the eight groups reported that brokers did not recommend the cheapest loan option in more than 20% of cases. ASIC was clear that a higher-cost loan can be the right recommendation — valuable loan features were the example given — but it said some brokers lacked clear evidence supporting that choice, and that licensees should monitor recommendations more closely where a cheaper option was available, with stronger evidence expected as the cost difference grows.

Put the two developments side by side. Lenders are narrowing the conditions under which their sharpest offers apply. ASIC is asking for better evidence when the recommended loan is not the cheapest available. The intersection is a file note problem that is easy to fix now and awkward to reconstruct later: if the cheapest option on your software was not accessible to this client — because of LVR band, loan size, security, income type, servicing or turnaround — the reason belongs in the file at the time, not in your memory.

This is general information rather than compliance advice. REP 840 is addressed primarily to licensees, and how its findings translate into your own process sits with your aggregator or licensee’s compliance team.

Retention is the other half of a proprietary build

Branch lenders, extended trading hours and trained mortgage sales forces are usually discussed as acquisition investments. They are also retention infrastructure. A bank with more lenders in more branches has more capacity to call a client at the point of a rate review, a fixed-rate expiry or a refinance enquiry — including clients that broker channel originated.

Westpac’s own pack lists lead quality optimisation as a proprietary focus area. NAB’s lists investment in frontline bankers. Neither is aimed at brokers, but both increase the chance that a client of yours has a conversation with their lender before they have one with you.

The defence is the one brokers already know and often defer: owning the review cycle. If your client’s next rate conversation is initiated by you, the bank’s additional branch capacity is irrelevant to your book. If it is initiated by the bank, your retention outcome is decided by someone else’s script.

What to review this week

• Pull your last 20 submissions and sort them by lender. If one or two lenders carry most of your flow, write down what would happen to your pipeline if either narrowed its appetite by LVR band or loan size.
• Map your panel by deal shape, not by rate. For each of the three or four file types you write most — high-LVR purchase, low-LVR refinance, self-employed, investment — list the lenders you would place it with first, second and third today.
• Check the conditions attached to your three sharpest advertised rates. Confirm the LVR band, minimum and maximum loan size and borrower type with the lender or your aggregator before you quote them again.
• Audit five recent files where you did not recommend the cheapest option. Does the file state, in its own words, why the cheaper option was not suitable or not available to that client?
• Identify the clients most exposed to a lender-initiated conversation. Fixed-rate expiries, loans written at the top of an LVR band and clients who have not heard from you in twelve months.
• Book the review calls you have been deferring. The proprietary investment the majors described is multi-year; the retention window on an individual client is weeks.
• Give your support team the one-page version. A panel-by-deal-shape sheet is more useful to a loan processor than a rate table.

A scenario worth working through

Consider two files you might write in the same week. The first is an owner-occupier purchase at 90% LVR with lender’s mortgage insurance, $620,000 loan, PAYG income, settlement in five weeks. The second is a $1.3 million refinance at 58% LVR, dual PAYG income, no time pressure.

On a rate table these may well point to the same lender. On a segmented panel they do not behave the same way at all. The second file sits squarely inside the profile NAB described wanting more of; the first sits outside it. One is a candidate for pricing discretion and attentive service; the other may be better served by a lender whose appetite at high LVR is genuine. Run both through your actual panel today, rather than assuming your default lender order still holds. Verify current rates, policy and promotions with the lender or your aggregator before relying on any of it — this scenario is illustrative, not a pricing guide.

What to watch next

Three things will tell you whether FY26 was a turning point or a cycle. First, whether NAB’s narrowed broker appetite shows up as declines and repricing in your own submissions over the next two quarters, or stays a statement on an earnings call. Second, whether ANZ’s branch lender build changes its broker share, given that share rose while the investment was being announced. Third, system credit growth: The Adviser reported NAB forecasting Australian system housing credit growth of about 2.6% in FY2026–27, against about 7% in FY2025–26. In a market growing at 7%, banks can grow both channels. In a market growing at 2.6%, channel share becomes a zero-sum contest — and that is the environment in which a proprietary investment starts to cost brokers volume rather than simply running alongside them.

The strategic takeaway

Brokers won the distribution argument years ago, and a record 81.6% share says the structural position is secure. What FY26 shows is that the argument has moved. The question is no longer whether banks will use the channel, but which files they want from it, from whom, and at what price — and the brokers who can answer that lender by lender, deal shape by deal shape, will quote more accurately, convert more reliably and evidence their recommendations better than brokers still working from a rate table.

That is a research task, not a defensive one. It costs an afternoon with your own submission data and three phone calls to your BDMs. In a year where three of the four majors have told you, in their own documents, what they are buying, declining to read it is the expensive option.