What a Safe Withdrawal Rate Means With an Age Pension
The 4% rule came out of US research on US market history, for a retiree with no means-tested state pension underneath them. Australia has one, and it works as a floor that rises exactly when a portfolio falls. That does not make the rule wrong; it makes it the answer to a different question from the one an Australian household is asking.
- The answer: A 4% initial withdrawal indexed to inflation was derived from US data for a 30-year horizon with no state pension. It is a starting point, not a rule.
- The trap: Applying a fixed percentage ignores the Age Pension, which rises as assets fall. An Australian retiree's worst case is materially better than the research assumes.
- The recommendation: Model the pension alongside the portfolio rather than treating the portfolio as the whole plan. The two move in opposite directions, which is the point.
Where the AI summary above gets this wrong
"The 4% rule says you can withdraw 4% of your retirement savings each year and never run out."
That's surface-true. Here's what it misses:
- It was never a promise of never running out — Bengen's work identified the worst historical starting year for a 30-year horizon in US data. It is a backtest of one market's history, not a guarantee about any future.
- It assumes no means-tested safety net — An Australian retiree drawing down a portfolio becomes eligible for more Age Pension as the assets fall, so income does not track the portfolio to zero.
- It assumes constant real spending for 30 years — Household spending in retirement falls in the later decades for most people, which makes a constant real withdrawal a conservative assumption on top of an already conservative starting point.
Take a couple retiring at 66 with $850,000 between them — a composite of a household right at the point where the question bites. A 4% withdrawal gives them $34,000 a year. Whether that is sustainable depends far more on the Age Pension behind it than on the percentage in front of it.
01 Where the number came from
The 4% figure originates in work on US market data that asked a narrow question: across every historical 30-year starting point, what initial withdrawal rate, indexed to inflation, would have survived the worst of them? The answer for a stock-and-bond portfolio was a little above 4%.
The Trinity study extended it across different portfolio mixes and payout periods, reporting success rates rather than a single safe figure. Both are backtests of one country's market history over one period, and both are honest about being exactly that.
Neither was written as advice, and neither claims a guarantee. The rule became a rule in the retelling, and the retelling dropped the assumptions that made the number meaningful.
The most important of those assumptions, for an Australian reader, is that the retiree has no other source of income. That was reasonable in the US context of the research and it is not reasonable here.
02 What the Age Pension does to the problem
The Age Pension is means tested, so it rises as assessable assets fall. A retiree whose portfolio has a bad decade becomes eligible for a larger payment precisely when they need it, which is the opposite of how sequence risk works on a portfolio alone.
That converts the failure mode. Without a pension, running out means income goes to zero. With one, running down a portfolio means income converges towards the full Age Pension, which is a real and indexed floor rather than nothing.
The practical consequence is that the question changes from 'will I run out?' to 'what standard of living am I defending, and for how long?'. The first has a percentage as its answer; the second has a number of dollars above the pension.
The taper is what does the work, and it is steep. Every $1,000 of assets below the threshold adds $78 a year of pension, as set out in the assets taper reference — an implicit return on spending down that no portfolio matches.
03
| The 4% rule as usually stated | An Australian household | |
|---|---|---|
| Income floor if the portfolio fails | None | The full Age Pension, indexed |
| Behaviour as assets fall | Income falls with the portfolio | Pension rises as assets fall |
| Horizon | 30 years, fixed | Unknown, and longer for a couple than for either partner |
| Spending pattern | Constant in real terms | Falls through the later decades for most households |
| Tax on withdrawals | Assumed | Nil from a taxed fund after 60 |
| What failure looks like | Zero income | Living on the Age Pension |
04 What to use instead of a fixed percentage
Start from the spending you actually need, split into a floor you will not compromise and a discretionary layer you would flex. The floor is the part the Age Pension and any guaranteed income should cover; the discretionary layer is what the portfolio funds.
Then check the withdrawal that produces against the balance, and against the age at which the Age Pension starts if you are retiring before 67. The years before pension age are the expensive ones, because the portfolio is carrying the whole load.
The worked example below applies a withdrawal rate to a balance and shows how long it lasts at a given return, and then shows what happens when a pension floor is added underneath. The second number is the one that reflects an Australian retirement.
None of this argues for spending more carelessly. It argues that the constraint is the standard of living you are defending rather than the risk of destitution, and those two produce different plans.
The split also tells you what to hold where. A floor that must not fail is not a job for growth assets, and a discretionary layer you are willing to flex by a fifth in a bad year is not a job for cash. Most portfolios that feel wrong in retirement are wrong because one pool is being asked to do both, and the mix that results satisfies neither requirement.
Shows: how long a balance lasts at a given withdrawal and return, and what total income a household has once an Age Pension floor is added underneath it. Ignores: the means test, which would raise the pension as the balance falls, inflation on both the withdrawal and the pension, sequence of returns, and tax.
05 Where the research still applies
Sequence-of-returns risk is real regardless of the safety net. A poor first decade does permanent damage because withdrawals sell units at low prices, and no amount of average return afterwards recovers the units that were sold — the mechanism is set out in the market downturn post.
The finding that a higher equity allocation improved outcomes across long horizons also survives translation. Portfolios held too conservatively fail more often over 30 years than portfolios with meaningful growth exposure, because inflation is the slower risk and it does not stop.
And the observation that flexible spending dramatically improves outcomes holds everywhere. A household that reduces withdrawals after a bad year survives scenarios that break a household committed to a fixed real amount.
What does not survive translation is the number itself. A percentage derived from one country's history, for a retiree with no pension, is not the answer to a means-tested question.
Source: Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable
06 The longevity assumption underneath it
A 30-year horizon from 65 takes you to 95. For a single person that is a conservative planning age; for a couple it is closer to the middle, because the relevant question is when the second of the two dies rather than the first.
The ABS life tables are the right source for the base expectation, and the number that matters is the expectation conditional on having already reached 65 rather than life expectancy at birth. The two differ by several years and the second is the one usually quoted.
Planning to a fixed age has the structural weakness that being wrong is asymmetric: dying earlier than planned leaves an estate, and living longer than planned leaves a problem. That asymmetry is the argument for a floor rather than for a lower withdrawal rate, and it is worked through in the longevity risk post.
It is commonly assumed that a longer horizon simply means a smaller safe percentage. It is not that simple — beyond about 30 years the sustainable rate flattens out, because a portfolio that survives three decades of withdrawals has usually grown enough to survive indefinitely.
07 What I would actually do
Work out the Age Pension you would receive at three balances: today's, half of it, and a quarter. Those three numbers describe the floor under every scenario you are worried about, and most households have never calculated them.
Then set the withdrawal from the spending, not from a percentage. If the resulting rate is above 5% of the balance in the years before Age Pension age, that is worth knowing and is frequently fine, because the pension starts later and the portfolio only has to bridge.
Then check what happens at 75 and at 85, because the minimum drawdown factor rises at both and eventually forces withdrawals larger than your spending. That is not a problem, but it is a different one — the money leaves the untaxed environment whether or not you have a use for it.
Review the rate after any year in which the portfolio falls materially, and be willing to defer the discretionary layer for a year. That single behaviour improves outcomes more than any choice of starting percentage.
And do not solve for never running out. Solve for never dropping below the standard of living you would find unacceptable — which, for most Australian households, the Age Pension already partly does.
If the answer that comes back is that you can afford considerably more than you are spending, that is a result rather than an error in the model. Australians draw down conservatively and a large share of superannuation is still unspent at death, which the Retirement Income Review recorded as a systemic feature rather than a set of individual choices.
The 4% rule answers an American question, and the American question is 'how do I avoid destitution with no state pension?'. Australia's question is 'how much above the Age Pension can I defend, and for how long?'. Those produce different plans, and the second one is far less frightening once you have actually calculated the floor.
FAQ
What is a safe withdrawal rate in Australia?
There is no single figure, because the Age Pension rises as your assets fall. Start from the spending you need above the pension rather than from a percentage of the balance, and check the rate separately for the years before Age Pension age.
Does the 4% rule work in Australia?
It answers a different question. It was derived from US market history for a retiree with no means-tested state pension, so its failure case — income falling to zero — does not exist here.
How long will my savings last?
That is set by the withdrawal, the return, and the sequence in which the returns arrive. The worked example above gives the constant-return answer; a bad first decade produces a materially shorter one for the same average return.
Should I plan to age 90 or age 100?
For a couple, plan to the second death rather than the first, and use the life expectancy conditional on having reached 65 rather than at birth. Beyond about 30 years the sustainable withdrawal rate flattens out, so the extra years cost less than people expect.
Does spending really fall in later retirement?
For most households it does, through the transition from active to less active years, though health and aged care costs can rise sharply at the end. A constant real withdrawal assumption is therefore conservative in the middle and optimistic at the very end.
What matters more than the withdrawal percentage?
Flexibility. A household willing to defer discretionary spending after a bad year survives scenarios that break a household committed to a fixed real amount, and that behaviour improves outcomes more than any choice of starting rate.
Sources
Regulator references
- ASIC Moneysmart — Retirement income · ASIC Moneysmart · 2026The sources of retirement income in Australia and how they combine.Last verified: 2026-09-07
- Services Australia — How much Age Pension you can get · Services Australia · 2026The current payment rates, and the March and September indexation that moves them.Last verified: 2026-09-07
- Services Australia — Assets test for Age Pension · Services Australia · 2026The assets test: which assets count, the thresholds, and the taper that reduces the payment.Last verified: 2026-09-07
- Australian Bureau of Statistics — Life expectancy · Australian Bureau of Statistics · 2026The ABS life expectancy series, used for the planning horizon.Last verified: 2026-09-07
Research
- Determining Withdrawal Rates Using Historical Data · Journal of Financial Planning · 1994origin of the 4% rule and the SafeMax conceptLast verified: 2026-09-07
- Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable · AAII Journal · 1998the Trinity study: withdrawal rates backtested against 1926-1995 returns across 15- to 30-year payout periodsLast verified: 2026-09-07
- Money in Retirement: More Than Enough · Grattan Institute · 2018modelled replacement rates against the OECD's 70% benchmark, and which households actually fall shortLast verified: 2026-09-07
Calculator unit tests · the assertions this page's worked example is checked against, and their last result
Changelog
- 2026-09-07 — initial publish (new format)
See how this decision plays out across your 30-year projection
Model this choice against your real numbers — month by month, to age 90.
Join the Waitlist
Jordan Reeves
Founder of Talk Through Wealth. A software engineer for over a decade before turning to retirement planning, Jordan built the projection engine after watching family members get fragmented, country-by-country advice that never reconciled. He writes about retirement the way the engine computes it: month-by-month, lifetime-long, and skeptical of any rule of thumb that hasn't been run through the math.