Mortgage Complaints Ran From 8 to 39 per 1,000 Accounts Across Five Banks. RG 273.71(b) Lists ‘Service Levels of the Credit Provider’ as a Non-Cost Factor
For six years brokers have argued service levels on file with no public evidence behind the claim. ASIC built the evidence base in March. The first mortgage-specific read of it landed this week — and it cuts both ways.
The short version: A new analysis of ASIC’s internal dispute resolution data puts mortgage complaint incidence across five banks between roughly 8 and 39 complaints per 1,000 mortgage accounts. That is the first public, comparable number brokers have ever had for a factor RG 273 has treated as relevant since 2020 — and the same guidance that makes service levels relevant also requires that any claim about them be able to be substantiated.
Most brokers have written some version of this sentence into a file note: the client chose this lender partly because of service. Turnaround times, how the assessment team handles a messy self-employed file, whether a discharge gets actioned without three follow-up calls. It is a real reason, brokers have always known it matters, and ASIC’s best interests duty guidance has always allowed it.
What the claim has never had is a number behind it. Service was something you knew from your own last twenty files and what your BDM told you. It was a professional judgement, not an evidenced one.
That is changing, in two steps. In March, ASIC launched an internal dispute resolution (IDR) data dashboard that — in the regulator’s own words — lets users “compare the complaints reported by individual financial firms for the first time.” This month, the first mortgage-specific analysis of that data was reported, and it shows a spread across five banks wide enough that no broker should keep treating lender service as unmeasurable.
1. What the analysis actually found
The figures come from a Mortgage Complaint Incidence analysis by the firm FeedbackIQ, reported by The Adviser on 16 November 2026. The analysis draws on ASIC’s publicly reported IDR data and measures complaints against the number of mortgage accounts each bank holds, so that a large book does not automatically look worse than a small one. The reference period is the six months to 31 December 2025.
As reported, complaint incidence per 1,000 mortgage accounts — expressed as “at least” figures — ran as follows: NAB 39, ANZ 28, Westpac 15, CBA 12, and Bendigo and Adelaide Bank 8. On those numbers the highest rate is close to five times the lowest. The Adviser also reported that all five banks sat above a United Kingdom mortgage complaint benchmark, though the comparison basis was not detailed.
Direction of travel differed from level. Across the year to the second half of 2025, combined complaint volumes across the five banks were largely unchanged, but the components moved apart: ANZ up 23 per cent, NAB up 11 per cent, CBA down 28 per cent, Westpac down 3 per cent and Bendigo down 7 per cent. A bank can therefore be carrying a high rate and improving, or a low rate and drifting the wrong way. Those are different conversations.
These are figures from a single analysis reported by a single outlet, not an ASIC finding or an industry-agreed benchmark. Treat them as the first read of a newly public dataset rather than a settled league table.
A complaint is not a finding of wrongdoing. A higher complaint rate is not misconduct, and nothing in this data establishes that any bank did anything improper. FeedbackIQ founder Janine Lloyd made the point herself in reporting the analysis: differences in complaint incidence may reflect reporting practices, customer and product mix, or genuine differences in customer experience — and understanding which of those is driving a number is necessary before drawing conclusions from it.
A bank that records complaints diligently and resolves them properly can look worse on a raw rate than one that logs fewer of them. That is a known weakness of all self-reported complaints data, here and overseas.
2. The paragraph that has been sitting there since 2020
Here is why this is a compliance story and not just a lender-league-table story.
ASIC’s Regulatory Guide 273, Mortgage brokers: Best interests duty, sets out what brokers may weigh when deciding whether a product is in a consumer’s best interests. RG 273.57 is blunt that price is not the whole answer: “Cost is not the only matter relevant to whether recommending a product is in the consumer’s best interests.”
RG 273.71 then lists the non-cost features a consumer may derive value from. Sub-paragraph (a) covers the substantive ones — an offset account, a redraw facility, a package. Sub-paragraph (b) covers something else entirely: “matters that are more difficult to quantify, such as the service levels and policies of the credit provider.”
RG 273.73 reinforces it, listing among the factors a broker could take into account “the credit policy and risk appetite of the credit provider” at sub-paragraph (d), alongside interest rate, fees and charges at (e).
Read those together and the position is clear: lender service levels have been a legitimate best-interests consideration since the guide was published in 2020. The phrase ASIC chose — “more difficult to quantify” — was an accurate description of the world at the time. It is a less accurate description of the world now.
The useful way to frame this: RG 273.71(b) did not change. The evidence available to support a RG 273.71(b) argument did.
3. Why the same guidance raises the bar on you
Brokers reading the table will reasonably see an argument they can now use. It is worth sitting with the other edge of it first.
RG 273.75 says that where a product is recommended on the strength of a feature, “you should be able to make an overall assessment that the non-cost consideration has a realistic possibility of offering that individual consumer greater value.” The guide then quotes paragraph 3.25 of the Replacement Explanatory Memorandum, which says a broker “should not recommend a loan by prioritising factors that cannot be substantiated as delivering benefits to that particular consumer (such as the broker’s relationship with the [credit provider]), over factors and features which affect the cost of the product or are more relevant to the consumer.”
RG 273.77 is more pointed again. Dealing with situations where non-cost features are highly relevant — its example is a consumer who prioritises approval time because of an impending settlement — it says: “we expect that any claims of this nature you make will be evidence-based and able to be substantiated.”
That sentence has always been in the guide. For most of the last six years, a broker asked to substantiate a service-levels claim could fairly answer that no comparable public data existed, and point to their own experience. That answer is weaker in a world where ASIC publishes firm-level complaints data and third parties are now analysing it. It does not become wrong — your own file experience is still evidence, and arguably better evidence for a specific scenario than an aggregate rate. But “service was better” as a bare assertion, on a file where the recommended loan cost the client more, is a thinner note this month than it was last year.
RG 273.78 is worth re-reading in the same sitting: there is an increased risk of non-compliance “if your processes typically lead to a ‘one-size-fits-all’ outcome for consumers.” A service-levels rationale that appears in identical wording on every file you write to your most-used lender is the kind of pattern that paragraph describes.
4. Read the issue mix, not the ranking
The headline rate is the least useful part of this data for a broker. The breakdown by issue is where the practical signal sits, because the categories map onto very different client experiences — and onto very different amounts of unpaid work landing back on your phone.
As reported, the banks’ relatively stronger complaint areas differed. NAB showed a higher incidence of complaints about fees and costs, and about failures or delays in carrying out customer instructions. ANZ showed a higher incidence of service-related matters and of hardship and vulnerability issues. CBA’s largest gaps relative to peers were in commercial practices and policies, and interest. Westpac’s were general service delay and the management of complainant details. Bendigo’s were general service delay and other disclosure-related issues.
Translate those into broker consequences and they stop being equivalent:
- Failures or delays in carrying out customer instructions is the category that becomes your problem. A redraw not set up, an offset not linked after a refinance, a repayment type not switched, a discharge sitting unactioned. The client does not ring the bank’s call centre first. They ring you, because you are the person they have a relationship with.
- Hardship and vulnerability matters most for a client whose circumstances are likely to move — a single-income household, a recent business start, a borrower stretched on servicing. If that client may need to have a hard conversation with their lender in two years, how that lender handles those conversations is squarely a best-interests consideration under RG 273.71(b).
- Fees and costs complaints are a signal about the gap between what was disclosed and what the client experienced. That gap is where your own file documentation gets tested.
- General service delay is the one brokers already price in and the one clients forgive fastest once the loan has settled.
A broker placing a straightforward PAYG refinance for a client who will never contact the lender again should weight this data very lightly. A broker placing a construction loan, an SMSF file, a complex trust structure, or a borrower on thin servicing is placing a client who will interact with that lender repeatedly, under pressure, over years. That is the file where the issue mix earns its place in your reasoning.
5. Three reasons not to over-read the table
The reference period is behind us. The incidence figures cover the six months to 31 December 2025. Since then the cash rate has moved up to 4.85 per cent, its highest level since 2008, on a split Reserve Bank board vote in November. Hardship and fee complaints are rate-sensitive. A complaints picture drawn from a lower-rate period is not a reliable description of how a lender is handling stressed borrowers in late 2026.
Complaint counts measure reporting as much as performance. This is the caveat the analyst raised herself, and it is not a technicality. Banks differ in what they classify as a complaint, how early they log one, and how they treat an expression of dissatisfaction that is resolved on the first call. A rate difference can be a culture-of-recording difference.
Five banks is not the market. The analysis covers five. Your panel almost certainly includes non-banks, mutuals and specialist lenders that are not in this comparison at all, and for many of the files where service genuinely decides the outcome, those are the lenders in contention. Absence from the table is not a good score.
There is also a client-facing consequence worth getting ahead of. ASIC’s dashboard is public. A client who reads a news article about lender complaint rates can look up the bank you recommended. Commissioner Alan Kirkland said at launch that the dashboard “provides a bird’s-eye view of how the Australian financial sector handles complaints” and makes it easier to identify “the reasons complaints are lodged, increases or decreases in complaints handling times, and the sorts of products that attract the most complaints.” If you recommended a lender that screens poorly on a measure your client can now find, you want your reasoning on file before they ask — not improvised afterwards.
6. Putting it into your process this week
This does not need a new system. It needs about an hour, and a change to how one sentence gets written.
- Pull five recent files where you recommended something other than the cheapest option. Find the sentence that justifies it. Ask whether it names a specific feature, policy or service behaviour and ties it to something you know about this client — or whether it is a generic statement that would read identically on anyone’s file. RG 273.75 is the test: can you make an overall assessment that the non-cost consideration had a realistic possibility of greater value for that individual consumer?
- Replace the bare assertion with the specific one. “Better service” is not substantiated. “Client is self-employed with a 19 November settlement; this lender assesses one year of financials and has returned formal approvals on comparable files within X days, where the cheaper option requires two years” is. RG 273.77 asks for evidence-based and substantiable claims, and your own file history is legitimate evidence — if you actually record it.
- Start keeping your own service record. Four columns in a spreadsheet: lender, file type, days to formal approval, and post-settlement issues. Six months of your own data beats any published aggregate for justifying a specific recommendation, and it is the one dataset no competitor and no regulator can tell you is irrelevant to your client base.
- Re-read RG 273.112 against your panel. It says you must be satisfied the products you can access are sufficient to let you act in clients’ best interests, and that “periodically comparing them to those you can access and recommend, may help you to determine whether your panel and accreditations are sufficient.” Public complaints data is now one input into that periodic comparison. If the lenders that handle vulnerable or complex clients best are ones you are not accredited with, that is a panel question, not a file question.
- Check that your records keep up. RG 273.167 lists what records may include, among them file notes and records of conversations at (b) and “outputs of product and feature comparison tools” at (e). If your reasoning now rests partly on a service comparison, the comparison belongs on the file, not in your head.
- A reported analysis of ASIC’s IDR data puts mortgage complaint incidence across five banks between roughly 8 and 39 per 1,000 accounts for the six months to 31 December 2025. It is one firm’s analysis reported by one outlet, not a regulator finding.
- Complaint incidence is not evidence of wrongdoing, and the analyst herself cautioned that differences may reflect reporting practices or customer and product mix rather than customer experience.
- RG 273.71(b) has listed “the service levels and policies of the credit provider” as a relevant non-cost factor since 2020. What is new is public data capable of supporting such a claim.
- RG 273.75 and RG 273.77 cut the other way: a non-cost claim should be substantiable and tied to the individual consumer. The strongest evidence for that remains your own recorded file history.
- The issue breakdown is more useful to brokers than the ranking, because categories such as failures or delays in carrying out instructions are the ones that return to the broker as unpaid work.
7. What to watch next
Three things worth tracking. First, whether later reporting periods on ASIC’s dashboard show the same spread once higher-rate conditions wash through hardship complaints — the 2026 periods will be the more relevant ones for the files you are writing now. Second, whether aggregators begin surfacing complaints or service data inside lender comparison tools, which would turn this from an article into a field on your CRM. Third, whether non-banks and mutuals become visible in comparable analyses, since for many complex files those are the lenders that matter most.
Separately, trade reporting on ASIC’s 2026–27 banking sector supervisory agenda indicates it includes a review of how lenders monitor broker-originated lending, covering accreditation and re-accreditation, application quality and customer outcomes. A regulator looking at how lenders oversee brokers, and a public dataset showing how those same lenders handle their own customers’ complaints, are two halves of the same transparency shift.
The bottom line
The genuinely interesting thing about this data is not which bank came last. It is that a factor ASIC described in 2020 as “more difficult to quantify” is becoming quantifiable in public — and the guidance that permits brokers to rely on it is the same guidance that asks them to substantiate it.
Brokers who already recommend on more than price gain something from this: an external reference point that supports reasoning they have been making on instinct. Brokers whose service-levels rationale is a sentence copied between files lose something: the excuse that nobody could check. Either way the response is the same, and it is not complicated. Write the specific reason, tie it to the client in front of you, keep your own record of how lenders actually perform on your files, and make sure the file says what you were thinking at the time.
Policy shifts, lender moves and compliance changes, read in the time between appointments.
More at The Broker Times →Questions brokers are asking
Can I put a lender’s complaint rate in a file note as a reason for my recommendation?
You can record a relevant, accurate consideration, and RG 273.71(b) treats the credit provider’s service levels and policies as relevant. Two cautions. Attribute it accurately — it is one firm’s analysis of ASIC-reported data for a period ending 31 December 2025, not a regulator ranking. And under RG 273.75 it should connect to why that matters for this client, not stand alone. Confirm the approach with your licensee before changing your file templates.
Does a higher complaint rate mean a lender has done something wrong?
No. A complaint is a customer raising an issue, not a determination against the firm. The analysis itself notes that differences may reflect reporting practices, customer and product mix, or genuine experience differences. Describing a lender’s complaint rate as evidence of misconduct would go well beyond what the data shows.
Am I now required to check complaints data before recommending a lender?
Nothing in RG 273 requires you to consult any particular dataset. RG 273.74 says ASIC does not expect brokers to comprehensively assess the separate value of each feature. The practical point is narrower: if you rely on service levels as a reason, RG 273.77 expects the claim to be evidence-based and substantiable. Your own recorded experience can do that work.
Where does the underlying data come from?
Financial firms report complaints received through their internal dispute resolution processes to ASIC. In March 2026 ASIC published an interactive IDR dashboard that, for the first time, allows comparison of complaints reported by individual firms, including issues, outcomes, handling times and monetary remedies. It also publishes guidance on interpreting the data and its methodology, which is worth reading before drawing conclusions.
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.

