Where the spending smile comes from
The term comes from David Blanchett’s 2014 Journal of Financial Planning paper, “Exploring the Retirement Consumption Puzzle.” He followed 591 retired households that answered all five RAND Consumption and Activities Mail Surveys from 2001 to 2009, matched to the Health and Retirement Study. Their inflation-adjusted spending fell by an average of 0.96% a year between ages 60 and 90. His title borrows the name economists give such a drop, the retirement consumption puzzle: life-cycle theory predicts that people try to keep consumption steady, the idea behind consumption smoothing.
The smile is in the rate of change, not in the level of spending. Declines were smallest for younger retirees, who were still traveling and active, and for the oldest, whose medical costs were rising, and largest in between. Blanchett stressed that the overall changes stayed negative; only their size varied. Many popular charts draw spending climbing back up in late life, but the underlying data show spending that falls and then falls more slowly.
Go-go, slow-go and no-go years
Planners often describe the same pattern as three phases of retirement. The ages in the list are only a guide: health, not birthdays, usually decides when one phase gives way to the next, and a couple can be in different phases at once.
The late-life phase carries the largest cost risk. According to the federal Administration for Community Living, someone turning 65 today has almost a 70% chance of needing some type of long-term care. Women need care longer on average (3.7 years) than men (2.2 years), and 20% of people will need it for more than five years. Those costs arrive in lumps rather than as a smooth rise, which is why many plans set aside a separate care reserve or weigh long-term care insurance.
- Go-go years, roughly the 60s to mid-70s: travel, hobbies and projects; often the highest-spending stretch of retirement.
- Slow-go years, roughly the mid-70s to mid-80s: less travel and driving, more time at home; spending usually falls fastest here.
- No-go years, roughly the mid-80s on: health and care costs dominate, and total spending can level off or climb if paid care is needed.
What government spending data show
The Bureau of Labor Statistics Consumer Expenditure Survey points the same way across age groups. In 2024, households headed by someone aged 65 to 74 spent $65,354 on average, and households headed by someone 75 or older spent $55,834, about 15% less. Transportation fell by 40% between the two groups and restaurant meals by 13%, while health care rose slightly in dollars and from 11.8% to 14.2% of the budget.
Two cautions apply. These are different households in the same year, not the same households followed over time, and the older group had lower incomes on average, so some of the gap reflects smaller budgets rather than choice. Prices matter too: Blanchett found that the basket retirees buy tends to rise faster than general inflation, largely because of medical care. BLS publishes a research price index for Americans 62 and older, the R-CPI-E, but warns that conclusions drawn from it should be treated as tentative.
How the smile changes retirement math
Most withdrawal research, including the studies behind the 4% rule, assumes spending rises with inflation every year. If real spending instead drifts down, the same savings can support more spending early on, and a flat income replacement ratio overstates later needs. Blanchett tested this with a Monte Carlo simulation of a 40% stock, 60% bond portfolio earning 3% a year after inflation. A 4% starting withdrawal lasted 30 years in 73.3% of trials when spending rose with inflation, and in 86.0% of trials when spending followed his curve for a $50,000 budget.
For a 65-year-old couple planning over their joint lifetimes, he found that a 5% starting withdrawal on the curve had about the same success rate as 4% with inflation-linked spending over 30 years. That is the same as needing 20% less savings: $800,000 instead of $1 million to fund $40,000 a year. His equation, shown in the formula box, was deliberately tilted up to leave room for future medical costs, so it is gentler than the raw data. Results like these depend on the return assumptions used, so treat them as a sense of scale rather than a new safe withdrawal rate.
Risks of planning on a spending smile
The smile is an average, and averages are a weak guide to one household’s future retirement income needs. The safest way to use it is selectively: model a decline only in the parts of the budget that are truly discretionary, keep essentials rising with prices, and give care its own line. Comparing a flat plan with a smile-shaped one, and pairing either with a dynamic spending rule, shows how much margin the assumption is really buying.
- Averages hide wide variation: Blanchett’s low-spending, high-net-worth households raised spending from 65 to 75, while high-spending, low-net-worth households cut sharply.
- Some of the decline may be forced: the data cannot fully separate retirees who chose to spend less from those who had to.
- Care costs are lumpy and can be large, so a smooth curve understates the risk in the no-go years.
- Living longer than planned adds years of spending, the heart of longevity risk.
- Using the smile to justify saving less leaves little room if you turn out to be an above-average spender.
Illustrative numbers
Flat vs. smile-shaped spending from 65 to 94, at today’s prices
- ΔAS
- Annual change in inflation-adjusted spending, as a decimal (−0.01 = −1%)
- Age
- The retiree’s age that year
- ExpTar
- Target annual after-tax spending in dollars
- ln
- Natural logarithm
For a $50,000 target, Blanchett’s 2014 equation gives about 0% at 65 and −1.4% near 78, and turns slightly positive after about age 91.
Flat plan: $60,000 a year for 30 years$1,800,000
Go-go years, 65–74: $60,000 a year$600,000
Slow-go years, 75–84: $50,000 a year$500,000
No-go years, 85–94: $55,000 a year, including care$550,000
Smile-shaped total$1,650,000
Difference from the flat plan$150,000, about 8%
The smile-shaped budget spends about 8% less over 30 years before any investment growth. The rise after 85 is a care budget layered on top of the average pattern, not part of Blanchett’s averages. Because the savings come in the middle decade and care costs arrive later, the effect on the portfolio today is smaller than the headline cut suggests.
At a glance
Average yearly spending by age of household head, BLS Consumer Expenditure Survey, 2024
| Category | Age 55–64 | Age 65–74 | Age 75+ |
|---|---|---|---|
| Total spending | $84,946 | $65,354 | $55,834 |
| Health care (share of spending) | $6,711 (7.9%) | $7,715 (11.8%) | $7,918 (14.2%) |
| Transportation | $15,085 | $11,414 | $6,855 |
| Food away from home | $4,040 | $2,841 | $2,472 |
| Entertainment | $3,706 | $3,122 | $2,888 |
| Housing share of spending | 31.8% | 34.2% | 39.4% |
| Income before taxes | $121,571 | $75,460 | $56,028 |
Put it in your plan
Spending Smile in MoneyWhatIf
A MoneyWhatIf spending card can follow inflation, rise or fall relative to it, or trace a custom curve, and separate date ranges can model distinct phases, such as travel spending that falls later in retirement. That lets you replace a flat line with a phased budget: a travel card that ends in your late 70s, a core budget that trails inflation slightly, and long-term care costs with their own timing and duration. Try the change in What-If to see it drawn against the dashed line of your current plan.
Common questions
Spending Smile FAQs
Is the retirement spending smile real?
The broad pattern is well supported: Blanchett’s panel data and BLS survey data both show older retirees spending less, after inflation, than younger ones. What is less certain is how much applies to you. The figures are averages, some of the decline reflects retirees who had to cut back, and health and long-term care costs can reverse the trend for a particular household.
When does retirement spending start to decline?
In Blanchett’s equation for a $50,000 budget, real spending is roughly flat around 65, declines about 1.3% to 1.4% a year from the mid-70s to about 80, and the decline then eases. BLS data show the same direction across age groups: households headed by someone 75 or older spent about 15% less in 2024 than those aged 65 to 74.
Why does spending go down in retirement?
Mostly because retirees do less that costs money as they age. In 2024 BLS data, households headed by someone 75 or older spent 40% less on transportation than those aged 65 to 74, and 13% less on restaurant meals. Part of the decline is not a choice: Blanchett found steep cuts among households spending more than their wealth could support. Health care is the main category that keeps rising.
Should I plan for my spending to go down in retirement?
Many planners keep a flat, inflation-adjusted budget as the base case because it is conservative, then test a smile-shaped budget as an alternative. A middle path is to hold essentials level, let travel and other discretionary lines taper in your late 70s, and add a separate budget for care. Revisit it as your health and plans become clearer; see retirement planning for the wider process.