Wages vs mortgage rates: how to read the serviceability gap

Three percent wage growth doesn't mean serviceability is improving. It depends entirely on what mortgage rates did in the same year. A pay rise that lands the same quarter your bank lifts your variable rate by 1.5 percentage points is not a pay rise at all in housing terms; it's a treadmill.

Two publicly published series tell you almost everything you need about which way the treadmill is moving: the ABS Wage Price Index (catalogue 6345.0) and the RBA's F6 Housing Lending Rates table. Read together, the spread between their year-on-year movements is the cleanest single read on Australian housing serviceability you can build for free.

Why one series isn't enough

Wages in isolation tell you what households earn. Mortgage rates in isolation tell you what debt costs. Neither answers the question buyers and refinancers actually care about, which is whether the gap between income and the cost of debt is widening or closing.

A wage print of +3.0% year-on-year sounds reassuring on the evening news. Pair it with a mortgage rate that rose 1.5 percentage points over the same year, and the household with a $700,000 loan is roughly $9,000 a year worse off in repayments before the wage rise even hits the bank account. The headline wage number doesn't lie, but it answers the wrong question if you stop reading there.

The two-series read is the standard frame in major-bank quarterly economist notes and the recurring serviceability commentary in the RBA Financial Stability Review. CoreLogic's monthly release also leans on it when describing affordability trends. None of those publications invent it; they all read the same two ABS and RBA tables.

What the ABS Wage Price Index actually measures

The Wage Price Index, ABS 6345.0, measures the change in hourly rates of pay for a fixed basket of jobs. It's published quarterly, with state and industry splits, roughly six weeks after the reference quarter. The number you see cited as "Australian wages grew 3.2%" is the all-industries, total hourly rates of pay including bonuses, seasonally adjusted, year-ended.

Through the post-2010 era the WPI has spent most of its life in a 2 to 4% YoY band. It pushed above 4% briefly during the 2022-23 tightness, then settled back into the 3 to 3.5% range through 2025 as the labour market loosened from extreme tight to merely tight.

What the WPI doesn't capture is at least as important as what it does. Bonuses paid as one-off discretionary amounts, sales commissions, equity comp at listed and private firms, and the wage premium people earn by switching jobs all sit largely outside the index. Households whose income depends heavily on those components see a bigger income swing than the WPI implies, in either direction. For a buyer running a borrowing-power calculation on a base salary plus variable comp, the WPI is a useful national anchor but a poor substitute for the household's own recent payslip history.

What the RBA F6 table actually measures

F6 Housing Lending Rates is the RBA's monthly statistical table covering the average interest rate paid on outstanding and new owner-occupier and investor housing loans. The series most worth tracking for serviceability is the outstanding-loans, owner-occupier, variable-rate row.

Outstanding-loans matters more than new-loans for one reason: most Australian households are paying the rate on the loan they already have, not the rate a bank would offer them today. New-loan rates move first when the cycle turns, but the outstanding-loans series is what shows up in actual household budgets. When the RBA cuts and banks pass the move through, F6 outstanding starts trending down within a month or two; the new-loans series moved earlier but matters for fewer people in any given month.

F6 is national. There is no state-level cut of the table. For a household in Perth or Hobart, the WPI state print is relevant on the income side, but the rate side comes from a national average. That mismatch is one of the limits worth keeping in mind.

The spread, with worked numbers

The serviceability spread is a one-line calculation: WPI YoY minus the YoY change in F6 outstanding-loans rate. Positive means serviceability is easing; negative means it's tightening; near zero means the two effects roughly cancel.

Three scenarios that show the range.

• Mild tightening cycle. WPI YoY +3.0%, F6 outstanding YoY +1.5pp. Spread is +1.5pp on paper, but the 1.5pp rate move on a $700k loan is roughly $9,000 a year of extra interest, while a 3% rise on a $90k household income is $2,700 of extra take-home before tax. The headline spread looks positive, but the dollars run negative for the leveraged household. Lesson: the percentage-point spread flatters serviceability when loan size is large relative to income.
• Easing cycle. WPI YoY +3.0%, F6 outstanding YoY -10pp on the rate level (i.e. the average rate fell from around 6.0% to around 5.4%, a roughly 10% relative move). Expressed as percentage points the rate fell about 0.6pp. Even on the more conservative percentage-point read, the spread is +3.6pp. Combined with the wage rise, the leveraged household is several thousand dollars a year better off. This is the mid-cycle-cuts regime.
• Sharp tightening, weak wages. WPI YoY +2.5%, F6 outstanding YoY +5pp on the level (rates went from around 3% to around 4.5%, roughly 1.5pp absolute). The spread is -2.5pp on percentage points and the leveraged household is sharply worse off. This is the 2022-23 regime that produced the mortgage-stress headlines.

Two things to note. First, expressing the rate move in percentage points (the 1.5pp form) is cleaner for household budgets; expressing it as a percentage change in the rate level (the -10% form) is what some commentators reach for to make the move sound bigger. Both are correct; the percentage-point version is the one to use for serviceability arithmetic.

Second, the spread is approximate, not literal. A spread of zero doesn't mean serviceability is unchanged; it means the two series cancel at the headline level, and your own position depends on loan size, fixed/variable mix, and how much of your income shows up in the WPI's definition.

Worked example: dollars on a real household

A household with a $90,000 single income and a $700,000 owner-occupier variable mortgage at 6.0%. Monthly P&I repayment on a 30-year term sits around $4,196. Annual repayments: roughly $50,350.

The RBA cuts and the bank passes through 1pp. The contracted rate falls to 5.0%. New monthly P&I on the same balance and term: around $3,758. The repayment drop is roughly $438 a month, or about $5,260 a year. You can confirm the new repayment for your own loan size and rate on the mortgage calculator, and check what the lower repayment does to maximum borrow on the borrowing power estimator.

Layer a 3% wage print on top. The $90,000 income lifts to $92,700, an extra $2,700 a year before tax, roughly $1,900 after tax at the marginal rate that household sits on. Add the rate-cut saving and the repayment-cut saving: combined, the household has roughly $7,000 to $8,000 a year of fresh budget room. That is the dollar shape of a +3% wage print coinciding with a 1pp rate cut, which is the kind of spread the May 2026-onward easing cycle has been producing.

Run the same exercise the other way. Same household, same $700k loan, but rates go up 1pp instead of down. Monthly repayment lifts about $460, the household is roughly $5,500 a year worse off in repayments, the wage rise covers a third of the gap, and net budget room shrinks by $3,000 to $4,000. That arithmetic is the mirror image of the easing case, and it's why the same WPI print can be a relief or a squeeze depending on which direction the F6 row is moving.

Where the bank economists land

Major-bank quarterly notes consistently use this spread when they describe the household serviceability cycle. The framing is usually some version of "real wage growth has resumed" (when WPI runs above CPI) followed by "and the easing in housing lending rates is now adding to disposable income at the household level" (when F6 is trending down). When wages are running below the rate rise, the same notes flip to "mortgage stress indicators" and lean on arrears data from APRA's quarterly statistics.

The RBA Financial Stability Review revisits the same gap every six months, framed around the share of households whose scheduled mortgage payments have moved past 30% of disposable income. That share is mechanically driven by the WPI minus F6 spread plus the level of the loan-to-income ratio at origination.

Where the spread breaks down

State-level wages, national rates. WA wages in 2024-25 ran a full percentage point ahead of national WPI at times, driven by mining-sector pay outcomes. A WA household reading the national spread underestimates how much serviceability improved at home. A Tasmania household reading the same national spread overestimates it.

Investor rates aren't in the owner-occupier row. Investor mortgage rates typically run 25 to 40 basis points higher than owner-occupier and move on the same cycle but not always by the same amount. A landlord modelling cash flow needs the investor F6 cut, not the owner-occupier one.

Fixed-rate borrowers feel the cycle on a delay. Households that fixed for three years in 2021 at sub-2% rates rolled off into 5.5%-plus repricing through 2024, regardless of what F6 outstanding was doing in any given month. The national spread averaged across all loans masks that cliff-edge experience entirely.

And households who own their homes outright don't care about F6. About a third of Australian dwellings are owned without a mortgage. For those owners, the relevant series is property prices and rates on term deposits, not housing lending rates. The spread is a serviceability read; for non-mortgaged owners the question doesn't apply.

How to actually use it

Once a quarter, when the WPI prints, look up the most recent F6 monthly release. Compute the YoY change in WPI and the YoY change in the F6 outstanding-loans owner-occupier variable-rate row, in percentage points. Subtract. Note the sign and rough magnitude. Cross-reference with what the most recent RBA Statement on Monetary Policy says about household cash flows; the framing usually matches.

For a personal refinance or buy decision, run your specific loan size and rate through the lump-sum repayment calculator to see what a hypothetical rate move does to your own repayment. The national spread is context; the household arithmetic is the decision. The mechanics of how a cash-rate move actually reaches your mortgage rate, and how long that takes, are unpacked in the RBA cash-rate impact article, which pairs naturally with this one. For the related read on how lenders translate a serviceability picture into a loan offer, the borrowing power vs serviceability piece covers the mechanics.

On Burbfinder, the borrowing-power and mortgage calculators let you take the spread down to a number for your own household. The two-series read tells you which way the national tide is running; your own arithmetic tells you whether your boat is rising with it.