Fixed, Variable or Split in 2026? How to Decide When Nobody Knows What Rates Will Do

Rates

Every rate rise brings the same question, and it’s almost always framed the wrong way. People ask “will fixing save me money?” — which is really asking “can I predict rates better than the bank can?”

You can’t. Neither can I. Neither, on the evidence of the last five years, can the banks.

Here’s a better way to think about it.

Fixing is insurance, not a bet

When you fix, you’re not outsmarting your lender. Fixed rates are priced off the market’s expectation of where rates are going, and the lender has better information than you do. On average, over time, fixing costs slightly more than floating — that’s the premium you pay.

What you’re buying is certainty. A repayment you can put in a spreadsheet and rely on. That’s insurance, and insurance is worth paying for when the downside of being wrong is serious.

So the real question isn’t “will rates go up?” It’s: “how badly would I be hurt if they did?”

The state of play, July 2026

The RBA has lifted the cash rate from 3.60% to 4.35% across three hikes this year, and has retained a tightening bias. Forecasts for what comes next are genuinely divided — Westpac has been tipping further increases toward a peak near 4.85%; NAB, ANZ and CBA have been calling a hold. No major bank is forecasting a near-term cut.

That division is itself the useful information. When the professionals disagree this openly, the honest position is: this is uncertain, so manage the risk rather than trying to win the guess.

When fixing makes sense

Your budget has no slack. If another 0.50% would genuinely hurt, buy the certainty

You have a young family, a single income, or a recent income change — anything that makes a repayment shock harder to absorb

You’ve just bought at your absolute limit and need a couple of years to grow into the loan

You value sleeping at night more than optimising the last few thousand dollars. That’s not a soft reason. It’s a legitimate one

When fixing is a bad idea

You’re likely to sell within the fixed term. Break costs on a fixed loan can be brutal and are not capped

You want to make large extra repayments. Most fixed loans cap them

You want a full offset account. Fixed loans usually don’t offer one, or restrict it — and in a high-rate environment, an offset is doing serious work for you

You might refinance. Fixed loans make you sticky, which is exactly why lenders like them

The split loan — the option most people should look at

Fix a portion, leave the rest variable. Say 60% fixed, 40% variable.

You get: - Certainty on most of your repayment - An offset account working against the variable portion - The ability to make extra repayments on the variable portion - Partial benefit if rates fall, partial protection if they rise - Smaller break costs if life changes

It is not the optimal choice in either scenario. It is the choice that is never catastrophically wrong — which, when nobody can forecast reliably, is usually the right objective.

Frequently asked questions

What happens when my fixed rate expires? You roll onto the lender’s variable “revert rate”, which is often notably higher than what they’d offer a new customer. This is where lenders make a lot of money. Contact us three months before your fixed term ends, not after.

How much are break costs? They depend on how rates have moved since you fixed and how much of the term remains. They can run into many thousands of dollars and are not capped by law. Ask your lender for a figure before you assume.

Can I make extra repayments on a fixed loan? Usually only up to an annual cap, and often with penalties beyond it. Check your specific product.

Should I fix for two years or five? The longer the term, the more you’re paying for certainty, and the more your life is likely to change within it. Most borrowers who fix choose two to three years.

Not sure which structure fits your situation? We’ll look at your actual cash flow, your plans, and your tolerance for a shock — then compare across 40+ lenders.

Book a chat →