On 1 February 2026 APRA brought into force a portfolio-level cap on high debt-to-income (DTI) lending. Authorised deposit-taking institutions (ADIs) can originate no more than 20 per cent of new residential loans, by value, at a DTI ratio above 6 times the borrower's gross income. The rule sits alongside the existing 3 percentage-point serviceability buffer rather than replacing it. Both must be satisfied at origination. Three months in, the practical impact is becoming visible in the data lenders publish to the regulator and in broker-channel conversations across the panel.
The arithmetic of DTI is simple. Total household debt (the new loan plus any existing debts including HECS-HELP, credit card limits and personal loans) divided by total gross household income gives the DTI ratio. A couple earning $180,000 combined who borrow $1,200,000 are at 6.67x DTI. Same couple borrowing $1,000,000 are at 5.56x. Where households previously sat in the 6x to 7x band routinely (Sydney and Melbourne first-home buyers especially), lenders now have to triage which 20 per cent of their book sits above 6x. The borrower-level question is who is in that 20 per cent and who is not.
Where the cap is binding
APRA's March quarterly residential lending data shows two of the four majors are now operating with new-business DTI mix at or near the 20 per cent cap. The other two have headroom to roughly 16 to 18 per cent. Customer-owned banks and most non-bank ADIs sit well below the cap, typically 8 to 14 per cent of new business above 6x DTI. The non-ADI specialist lenders (which are not directly captured by the rule but face indirect pressure through funding-line covenants) are operating broadly as before, with the highest concentration of 6x-plus business in the panel.
The binding effect at the major-bank level shows up as policy tightening rather than outright decline. Brokers in the partner network report that majors are increasingly asking for additional buffer on already-tight applications, requesting six months of bank statements where three was previously enough, and applying tougher haircuts to bonus and overtime income. None of this is a rule change; it is the credit team's response to the portfolio-level constraint.
Who is being squeezed
Three borrower profiles are most affected by the cap and the second-order tightening it has produced. First, single-income high-DTI buyers in capital-city markets where house prices remain in the high $700ks to mid $900ks. A $900,000 loan against $130,000 income is 6.92x DTI; the same applicant a year ago might have been approved by a major; today they are typically referred out of the major panel into customer-owned and non-bank lenders.
Second, self-employed and contractor borrowers whose income calculation already involves haircuts and add-backs. The DTI calculation uses pre-haircut income (gross), but the serviceability buffer applies to the lender's assessed income. The combination tends to put self-employed applicants closer to the 6x line on paper than they sit in lender models, and the major-bank credit teams are increasingly conservative on borderline files. Specialist self-employed lenders have not changed materially.
Third, investors with multiple existing properties. Existing investment debt is captured in the DTI calculation, regardless of rental income. An investor with $1.4 million in existing debt across two properties who wants to add a third has a starting DTI position before the new loan even computes. Major-bank investor channels are visibly tighter in the second quarter of 2026, with several reducing their high-DTI investor mix below their portfolio cap.
The 6x DTI cap is the most consequential macroprudential rule for residential lending since the 2017 interest-only restrictions. The bite is selective, but where it bites, it is decisive.
Where the cap is not binding
Lenders with headroom under the 20 per cent cap have actively used it as a competitive lever. Customer-owned banks (Bank Australia, P&N, Heritage), most regional banks, and a handful of non-bank ADIs are still writing business in the 6x to 7x DTI band where applicants otherwise qualify on serviceability. Broker channels are the primary distribution path for these lenders, which is why borrowers turned away by their bank often find an approval through a broker without changing their underlying application materially.
Non-ADI specialist lenders (Pepper, Resimac, Liberty and others) are not directly captured by the cap, though their funders typically impose similar constraints through wholesale agreements. For most prime DTI-stretched applicants, ADI customer-owned options are cheaper than specialist non-ADI options. For applicants with credit impairment plus high DTI, specialist lenders remain the workable path.
What to do this quarter
- Calculate your DTI before you apply. Total household debt (new loan plus existing debts plus card limits at 3 per cent monthly) divided by gross income.
- Reduce or close unused credit card limits and personal loans where possible. The borrowing-power gain is typically much larger than the inconvenience cost.
- If you are a self-employed applicant, prepare two years of tax returns plus the most recent BAS for any year-on-year revenue movement that needs explaining.
- If a major bank declines or asks for excessive buffer, ask a broker to test the same application against the customer-owned and non-bank panels before accepting the rejection.
- For investors, model the cumulative DTI position across all existing debt before agreeing to a new purchase. The trap is the third or fourth property, not the first.