The retirement spending smile describes a U-shaped pattern of spending across retirement: higher in the early "go-go" years, lower through the middle, then higher again late in life as health and care costs climb. If you're planning your drawdown strategy around a flat, ever-increasing spending line, you're likely overestimating what you'll need in your seventies and underestimating what you'll need in your late eighties.
The practical implication matters more than the shape itself. A U-curve changes how you think about a sustainable withdrawal rate, because a single fixed percentage assumes spending stays constant or grows with inflation every year, but it doesn't for most retirees. Early retirement tends to front-load discretionary spending on travel and experiences while the mortgage is gone and health is good. The middle years usually settle into a quieter, cheaper rhythm. Then late retirement brings a second peak, driven largely by healthcare, home support, and aged care rather than lifestyle choices.
This pattern was first formalised by financial researcher David Blanchett in 2014, and it's since been revisited, refined, and occasionally challenged by later research, including a 2026 SSRN working paper by Tharp that re-examines the cross-sectional evidence behind it. Investor education bodies like FINRA also frame retirement income planning around exactly this kind of phased thinking rather than a static number.
Before you touch your withdrawal settings, three things are worth doing straight away:
- Check whether your current budget assumes flat spending for 25 to 30 years, because that assumption rarely survives contact with reality.
- Start building a healthcare and aged-care reserve now, even if it feels premature in your fifties or early sixties.
- Model at least two withdrawal scenarios, one with front-loaded early spending and one with a smoother, conservative path, and compare how each affects the odds your money lasts.
Table of Contents
- What is the retirement spending smile, and how strong is the evidence?
- Why does retirement spending rise, fall, then rise again?
- How does the spending smile change withdrawal strategy?
- How do you build a retirement budget around the spending smile?
- When doesn't the spending smile apply?
- How do you model the spending smile in your own plan?
- Should you actually plan around the smile, or play it safe?
- Key takeaways on the retirement spending smile
- Sources
What is the retirement spending smile, and how strong is the evidence?
Blanchett's original 2014 research plotted real household spending across retirement and found it didn't decline in a straight line, nor did it rise steadily with inflation as many models assumed. Instead it dipped through the middle years before curving back up. That curve, once you graph it, looks unmistakably like a smile, hence the name.
The idea has held up reasonably well, though the strength of the evidence varies by data source and country. Here's a short timeline of how the thinking has developed:
- 2014 – Blanchett's original study establishes the U-shaped pattern using US retiree spending data, challenging the assumption of flat, inflation-adjusted withdrawals.
- Ongoing – Practitioner explainers, including one from the Indiana Public Retirement System, popularise the concept for everyday retirement planning.
- Australian data – Milliman's analysis of Australian retiree expenditure finds median couple spending falls by more than a third between ages 65 to 69 and 85 plus, a steeper decline than many US studies suggest, with discretionary spending dropping consistently while health costs rise then dip again after 80.
The Milliman findings are worth sitting with, because they're a genuinely different shape to the US pattern in places.
Australian retiree couples show median expenditure falling by roughly a third from their late sixties to their mid-eighties, with the steepest declines in discretionary categories, not health, which suggests the "smile" in Australia may be more of a gentle downward slope with a less pronounced late-life upturn than the US data implies.
The honest takeaway is that the U-shape is a well-supported general pattern, not a guaranteed personal outcome. Cross-sectional studies (comparing different retirees at one point in time) can overstate how much of the decline is genuine ageing effect versus generational spending habits, since people currently in their eighties grew up with different financial norms than today's sixty-year-olds. Longitudinal data, which tracks the same households over time, tends to show a less dramatic curve. Treat the smile as a strong planning heuristic, backed by FINRA-aligned investor education principles, rather than a formula to plug in unchanged.
Why does retirement spending rise, fall, then rise again?
Four separate forces combine to produce the U-shape, and understanding which ones apply to you is more useful than memorising the curve itself.
Behavioural drivers show up first. Many new retirees deliberately front-load spending on travel and big experiences because they're healthy, mobile, and finally free of work obligations. There's also a documented reluctance to draw down capital once the initial burst passes, partly from habit and partly from anxiety about outliving savings. Interestingly, longitudinal wellbeing research cited alongside Blanchett's work suggests retirees often report higher life satisfaction even as real spending falls in the middle years, which undercuts the assumption that less spending automatically means a worse retirement.
Lifecycle and cashflow drivers do a lot of the heavy lifting in the middle years. The mortgage is usually paid off, work-related costs (commuting, professional wardrobes, workplace lunches) disappear entirely, and household composition often simplifies as children become financially independent. Spending naturally contracts because the cost base contracts.
Health and care drivers dominate the late upturn. Medical costs, home support, and eventually aged care tend to climb from the late seventies onward, and these costs are far less discretionary than a European river cruise. This is the part of the curve retirees underestimate most often, because early retirement health tends to feel deceptively stable.
Economic drivers interact with all of the above. Sequencing risk, where poor market returns early in retirement do outsized damage to a portfolio, is at its most dangerous exactly when go-go spending is highest. Health inflation also tends to run hotter than general inflation, which quietly erodes a fixed late-life budget faster than a retiree expects.
Three broad profiles illustrate how this plays out:
- The go-go retiree (roughly ages 60 to 75): prioritises travel, hobbies, and family experiences; spends at or above pre-retirement levels for the first several years.
- The slow-go retiree (roughly mid-70s to early 80s): spending settles as physical activity naturally reduces; discretionary categories shrink faster than fixed costs.
- The no-go retiree (80s onward): spending shifts almost entirely toward health, home modifications, and care support, often rising sharply if aged care becomes necessary.
How does the spending smile change withdrawal strategy?
A flat safe withdrawal rate, the "take out 4% and adjust for inflation every year" approach, was never designed around a U-shaped spending curve. It assumes constant real spending, which is precisely the assumption the smile contradicts. If you're prepared to front-load spending and taper through the middle years, your effective starting withdrawal rate can often be higher than the traditional rule suggests, provided the drawdown eases back once the go-go years pass.
The trade-off is sequencing risk. Spending aggressively in the first five to ten years of retirement, right when a market downturn does the most permanent damage to a portfolio, is the single biggest threat to plans built around a spending smile. Industry guidance on moving beyond the 4% rule increasingly points toward flexible, phase-based withdrawal rules rather than one static percentage locked in at retirement.
Income sources tend to map naturally onto the three phases, and matching the right funding source to the right phase is where most of the planning value sits.
| Spending phase | Typical funding source | Planning action |
|---|---|---|
| Early (go-go) | Account-based pension drawdowns, cash reserves | Fund from a liquid, short-duration bucket to avoid selling growth assets in a downturn |
| Middle (slow-go) | Account-based pension, Age Pension entitlements | Rebuild the liquid bucket during lower-spending years while markets recover |
| Late (no-go) | Age Pension, annuities or guaranteed income, home equity | Hold a dedicated healthcare reserve; consider partial annuitisation for baseline cover |
Pro Tip: Fund the first three to five years of retirement spending from a short-duration, low-volatility bucket separate from your growth portfolio. This means a market downturn in year two doesn't force you to sell shares at a loss just to cover a holiday you'd already budgeted for.
How do you build a retirement budget around the spending smile?
Turning the concept into an actual plan is a five-step process, and none of the steps require sophisticated software to start.
- Record your current lifestyle spending by category, separating genuinely discretionary items (travel, entertainment, dining) from fixed costs (insurance, utilities, rates).
- Stress-test the early spike. Model what happens if you spend 20 to 30% more than baseline for the first five years, and check the effect on long-term portfolio survival.
- Set a dedicated healthcare and aged-care reserve. This should sit outside your everyday drawdown bucket and grow steadily rather than being raided for discretionary spending.
- Decide on flexible withdrawal rules, such as adjusting spending down in years following a market fall (a "guardrails" approach) rather than locking in a fixed percentage regardless of returns.
- Review and rebuild the plan every two to three years, because health, family circumstances, and market conditions all shift the shape of your own personal curve.
A simple three-column budget makes the phases concrete rather than abstract:
| Expense line | Early retirement | Mid retirement | Late retirement |
|---|---|---|---|
| Travel and leisure | High, often top spending category | Reduced, more modest trips | Minimal |
| Housing costs | Low if mortgage is cleared | Low, occasional maintenance spikes | May rise with modifications or downsizing |
| Healthcare and support | Low, routine costs only | Gradually rising | High, potentially including aged care |
| Everyday living | Similar to pre-retirement | Reduced as activity slows | Reduced discretionary, stable essentials |
Decade-by-decade budgeting is the approach most Australian retirement planners now recommend, precisely because it forces you to revisit assumptions rather than "set and forget" a single number for 30 years. Some retirees also use the go-go, slow-go, no-go framework alongside ASFA retirement standard figures to sanity-check whether their planned spending in each phase is realistic for their target lifestyle.
A bucket structure, where short-term spending sits in cash or term deposits, medium-term needs sit in balanced assets, and long-term reserves stay invested for growth, works well against this budget shape. Staged asset sales and partial annuitisation both help smooth the transition from the go-go bucket into the no-go reserve without forcing a sale of growth assets at the wrong time. If you're weighing up how home equity or an investment property fits into that later-life funding gap, it's worth reading how property can support a secure retirement before locking in a strategy.
Pro Tip: Convert only enough capital to guaranteed income to cover essential late-life expenses, not your entire late-retirement budget. Keeping a portion liquid preserves flexibility if aged-care costs turn out lower, or higher, than you projected.
When doesn't the spending smile apply?
The U-shape is a population-level pattern, and population averages hide enormous variation between individual households. Some researchers, including the 2026 SSRN revisit, argue the "smile" is weaker than originally presented once you strip out generational spending habits from genuine ageing effects, and that averaging across very different households flattens real diversity into a tidy curve that doesn't describe any single retiree particularly well.
A few data limitations are worth understanding before you lean too hard on the pattern:
- Withdrawals aren't the same as spending. Some retirees withdraw more than they spend for tax or estate-planning reasons, which distorts studies that use withdrawal data as a proxy for consumption.
- Household heterogeneity is large. A couple with a paid-off home and strong superannuation balance behaves very differently to a single retiree still renting.
- Health shocks are endogenous, meaning they're not random. Retirees who experience an early health decline spend differently to those who stay well into their eighties, and averaging the two groups together can mask both patterns.
- Country and cohort differences matter. The steeper Australian decline found by Milliman doesn't necessarily transfer to US or UK retirees with different healthcare systems and pension structures.
The smile is a useful default assumption, not a guarantee. Households facing ongoing high caregiving costs, an unstable housing situation, or genuinely low risk tolerance often see a flatter curve, or one that rises earlier than the textbook pattern suggests.
Three caveats worth keeping in mind whenever you apply this pattern to your own numbers: the curve describes an average household, not necessarily yours; a health event can compress decades of the pattern into a much shorter timeframe; and generational differences in spending habits mean today's sixty-five-year-olds may not follow the exact path of the retirees the original studies measured.
How do you model the spending smile in your own plan?
Turning the U-curve into a genuine forecast means running scenarios rather than trusting a single static projection. A handful of inputs matter more than the rest, and varying them one at a time shows you which assumptions your plan is actually sensitive to.
- Early discretionary spending index: how much higher than baseline you plan to spend in years one to five.
- Mid-life decline rate: how quickly discretionary spending eases back as activity naturally reduces.
- Late-life healthcare shock: a one-off or ongoing cost increase modelled from your mid-eighties onward.
- Longevity assumptions: modelling to age 90 versus age 95 changes the required reserve substantially.
- Returns volatility: particularly in the first decade, when sequencing risk does the most damage.
When you run a scenario, the outputs worth actually inspecting are the probability your portfolio survives to your planned age, the "ruin risk" in years where markets underperform early, and the spread of possible outcomes rather than just the average case. A model that only shows you the median result hides exactly the downside risk the spending smile is meant to help you manage.
Pro Tip: Ask any modelling tool or adviser to test a front-loaded spending scenario combined with 1% higher healthcare inflation from age 85 onward. That single combined stress test reveals more about plan fragility than a dozen milder variations.
Any credible modelling exercise should be transparent about its own limitations too. Look for tools that report sensitivity to different assumptions, distinguish clearly between withdrawals and actual consumption, and avoid presenting a single confident number as if retirement were predictable to the dollar. Platforms like AlphaIQ build this kind of scenario simulation directly into tax-aware retirement modelling, which makes it easier to test a front-loaded spending assumption against your actual superannuation balance and asset mix rather than a generic industry average.
Should you actually plan around the smile, or play it safe?
The spending smile is one of the more useful ideas in retirement planning precisely because it corrects a bad default, the assumption that spending stays flat for three decades, without replacing it with an equally rigid new rule. My view, after working through the evidence base, is that the pattern deserves to shape your budget structure but not your withdrawal formula. Use it to size your buckets and your healthcare reserve. Don't use it to justify spending more in year one simply because a graph says most people do.
Where I'd push back on the more enthusiastic takes on this idea is the assumption that the late-life upturn is purely a healthcare story you can insure away with a fixed reserve. Aged care costs are lumpy, unpredictable in timing, and can arrive a decade earlier than planned if a health event forces the issue. A reserve sized on average figures will undershoot for a meaningful share of households, which is exactly the heterogeneity problem the Tharp revisit points to.
The practical next step is straightforward: run two scenarios, one that front-loads spending in line with the go-go pattern and one that keeps spending flat and conservative, and see how far apart the outcomes actually land for your specific balance and age. If the gap is small, the smile probably isn't decision-critical for you. If it's large, that's your signal to build the bucket structure properly rather than relying on a single fixed withdrawal percentage for the next 30 years.
Key takeaways on the retirement spending smile
The retirement spending smile shows spending typically peaks early, dips mid-retirement, and rises again late, which means a flat withdrawal rate usually misallocates risk across your retirement timeline.
| Point | Details |
|---|---|
| U-shape is well documented | Blanchett's 2014 research and later revisits confirm higher early and late spending with a mid-retirement dip. |
| Australian data shows a steeper decline | Milliman finds median couple spending declines significantly from age 65–69 to 85+. |
| Sequencing risk peaks early | Front-loaded go-go spending coincides with the years market downturns cause the most lasting damage. |
| Buckets beat flat rules | Matching short-term, mid-term, and late-life funding sources to each phase manages risk better than one fixed rate. |
| The pattern isn't universal | Heterogeneity, health shocks, and cohort differences mean some households see a flatter or reversed curve. |
Three priority actions: model two withdrawal scenarios (front-loaded versus conservative) before finalising your drawdown rate; build a short-duration cash bucket to fund your first five retirement years; and set aside a dedicated, separately tracked healthcare and aged-care reserve well before you expect to need it.
Sources
- Something to Smile About: How the “Retirement Spending ...
- Analysis: Retirees’ spending falls faster than expected into old age
- Beyond the 4% rule – managing retirement spending | BT
- Finra
If you want to see how these dynamics play out against your own numbers, AlphaIQ's superannuation calculator lets you model front-loaded and conservative withdrawal scenarios side by side, using your actual balance, age, and asset mix rather than a generic industry average.
This article is general information, not a substitute for advice from a qualified financial advisor. Consult a qualified financial professional about your own circumstances before acting on anything here.