Lean FIRE vs Fat FIRE: The Tiers Are a Guess About Your Spending Floor. I'd Rather Price Mine.

Last Sunday I opened the three top-ranking pages on lean FIRE vs fat FIRE and wrote their numbers on the back of an envelope, to see whether they agreed with each other.

They don't. The Motley Fool puts fat FIRE at "annual expenses of $100,000 or more." ProjectionLab says $200,000-plus. On the multiple, MoneyCrashers uses 28.6×, the Fool uses 25× and Fidelity's 33× in the same article. Read two of those pages instead of three and you walk away with targets that differ by more than double.

Nobody is lying. ProjectionLab says on the page that there's no official threshold, which I appreciate. That's the tell. If the biggest publishers in the genre can't agree on the definition to within 2×, the tier was never a thing you choose. It's a thing that falls out of a forecast you haven't done yet.

Lean FIRE vs Fat FIRE Is a Result. Everyone Sells It as a Choice.

Every page I read ends the same way, with a paragraph inviting you to pick the path that fits your values. The Fool's third heading is literally "Which approach is right for you?" MoneyCrashers tells you that if living simply sounds like a sacrifice, lean FIRE probably isn't for you.

That's a personality quiz standing where a forecast should be.

Lean and fat aren't lifestyles. They're two bets about how much of your spending is permanent, and the bet gets settled by one piece of work almost nobody does: separating the spending you'd defend if your income stopped tomorrow from the spending you'd drop without noticing. Do that and the label is just a name for the answer. Skip it and you've multiplied a label by a number you found on a blog and called it a plan.

Two of the FIRE flavours escape this. Coast FIRE and barista FIRE describe how you get there: what you stop doing, what income you keep. They're mechanisms. Lean and fat describe how much you spend, which is the one input nobody has measured yet. The vocabulary looks like one family of words. It's two.

Four Times the Money Buys Under Twice the Spending

You'll find articles this year claiming retiree spending "converges" regardless of portfolio size. It doesn't converge. It scales far more slowly than wealth does, and the extra headroom mostly goes unspent. That's a weaker claim than the one I set out to make, and it's still enough to wreck the tier genre.

J.P. Morgan's 2026 Guide to Retirement is the best evidence I found, because it isn't a survey asking people what they spend. It's observed transactions from US Chase households, 2017 to 2025, sorted by investable wealth. The US households holding $250,000 to $750,000 peak at $77,060 a year and drift to $53,980 at the oldest ages. The ones holding $1m to $3m peak at $136,810 and drift to $88,980. Four times the wealth, about 1.8 times the spending at peak.

Data: J.P. Morgan Asset Management 2026 Guide to Retirement (observed US Chase household transactions; only the labelled peak and oldest-age values are plotted) and the BLS Consumer Expenditure Survey 2024.

The Bureau of Labor Statistics arrives from a different direction and says the same thing. In the 2024 Consumer Expenditure Survey, the bottom fifth of US households spends $35,046 a year and the top fifth spends $150,342. The top quintile has no upper income limit at all. Unlimited income, 4.3 times the spending.

David Blanchett's spending-smile work adds the sentence that carries both halves of my argument at once. Higher-net-worth households' real spending declines throughout retirement: their whole curve of year-on-year real changes sits below the zero line. Lower-net-worth households finish roughly where they started. The wealthy cut. The lean don't. Hold that for two sections.

The caveats, because I'd rather hand them over than have you find them: Chase customers only, wealth bands estimated by a third party, and cross-sectional, so a 90-year-old on $53,980 and a 60-year-old on $77,060 are different people rather than one life. All of that attacks the size of the gap. None of it touches the direction.

Lean FIRE's Real Risk Is a Promise, Not a Budget

Lean FIRE isn't fragile because $40,000 a year is uncomfortable. Plenty of households live well on a good deal less than that and I'm not going to lecture them.

It's fragile because it quietly promises that your floor never rises again.

There's a real number for that floor, and it's academic rather than bloggy. Raj Chetty and Adam Szeidl, in Econometrica, found more than 50% of the average US household's budget stays fixed under a moderate income shock such as unemployment. Not because people cling to a standard of living, but because leases, insurance, school places and the logistics of moving carry real adjustment costs. Half your budget is a contract, not a preference.

Three things raise that half without asking permission, and all three are measurable.

A child. The USDA priced a child born in 2015 at roughly $1,030 to $1,160 a month to age 17, in 2015 money, and 29% of that is housing. The Child Poverty Action Group, using Loughborough's Minimum Income Standard, priced a UK child in 2025 at about £1,160 a month for a couple. Two methods, two continents, near-identical magnitude.

Housing. US CPI rent of primary residence rose 33.1% between January 2020 and August 2026; the euro-area equivalent rose 15.9%. On a $1,500 rent that's an extra $496 a month, arriving without a recession and without anyone voting for it. At 25×, that renter's number was understated by about $149,000 the day they wrote it.

Data: US BLS CPI rent of primary residence (CUSR0000SEHA) and Eurostat HICP actual rentals paid by tenants (CP0410EZCCM086NEST), via FRED. Each series is indexed to its own January 2020 base.

Health. The one line that rises in share and in absolute terms while everything else falls. JPMorgan has health care going from 6.9% of a US household's spending at 35-44 up to 15.8% at 75-plus, and it carries the highest long-run category inflation in the whole US CPI series, 4.4% a year since 1982. And before anyone files that as an American problem, euro-area health prices rose 18.1% since January 2020 against 15.8% in the US. Slightly faster on my side of the Atlantic.

Now the arithmetic that made me write this. A permanent €400 a month rise in your floor, times twelve, times twenty-five, moves your target by €120,000, and nothing recovers that except saving it. A 20% market drop takes a bite out of your balance and leaves your target where it was. I want to be careful there, because it reads as "market risk is trivial" and it isn't — a drop early in retirement is sequence-of-returns risk and can end a plan by itself. The narrow point is that one event changes your destination and the other changes your route.

And the case that settles it for me happened to someone who was never remotely lean.

Sam Dogen retired in San Francisco at 34 in 2012 on a $3 million net worth, comfortably fat by every definition in this article's first paragraph. He and his wife hadn't planned on having children when they retired. They now have two, and he reckons educating both could run to $1.5 million. Then in 2023 he bought, in his own words, an expensive home he didn't need, and took a part-time job at a fintech startup. Not a bear market. Two children and a house. The label protected him from nothing, because what got him was a floor he'd never priced.

Fat FIRE's Real Cost Is Working Years You Spend on Optionality

So aim high and stop worrying? No, and this is where I annoy the other camp.

A fat target is usually not a plan to spend more. It's a plan to never have to decide, and most of that headroom never gets used: EBRI's 2024 survey of US retirees found 38% with a savings mindset and only 11% with a spending one. I've argued that case out in die with zero vs FIRE, so I'll add the half nobody quotes. Between 40% and 43% of middle- and high-asset US retirees had less than half their money left twenty-one or twenty-two years in. "Everybody hoards" isn't a claim I can make.

The currency you buy that headroom with is years at a desk, and those years have a shelf life. Eurostat puts healthy life years at birth at 62.8 for EU men and 63.3 for women, against life expectancies of 78.7 and 84.0. The average European's healthy stretch ends around 63, and on the US side JPMorgan has spending peaking at 45 to 54 and already falling by the late 60s. Two separate datasets, pointing the same way. The ability to enjoy money declines before the money does.

You buy the headroom with the good years, and the good years are the part that runs out.

That's the honest limit on my own argument, and I'd rather say it than have someone say it for me. If health fades that fast, some over-saving is cheap insurance. The exchange rate only turns terrible past the point where your floor is covered twice over. Before that, buying certainty is just sensible.

My own target belongs to neither tribe, which is probably why I find the tribes irritating. I want free Wednesdays and the option to teach maths and physics for a fraction of what I earn now. That's not a lean number and it's certainly not a beach in a fat one. It's a floor plus a deliberately chosen amount of headroom, and I can already buy one of those Wednesdays without waiting for a tier.

Three Columns, One Evening, No Quiz

Here's what replaces the values quiz. Pull twelve months of real transactions and sort every recurring line into three columns: would defend if my income stopped tomorrow, would drop without noticing, genuinely don't know. Twelve months, not one. The US BLS runs four quarterly interviews plus a two-week diary to measure exactly this, because one month can't see the insurance renewal, the dental work or the boiler.

Then the correction I'd make to my own instinct, courtesy of Chetty and Szeidl: don't sort by what you'd nobly defend, sort by what's under contract. The fixed half of a budget is fixed by adjustment costs, not virtue. So the useful question isn't "would I give this up?" but "what would it cost me, in money and weeks, to stop paying this?" Rank the defend column that way: contract and notice period first, then physical switching cost, then things with a person attached.

The third column is the interesting one and it's usually the biggest. Don't resolve it by guessing. Test it for a quarter and see what you actually did.

Then two sanity checks. Composition is strikingly similar on both sides of the Atlantic, with housing taking 33.4% of US spending and 32.9% of EU spending, but the split underneath changes your floor's price tag: Americans put 17.0% into transport against 10.9% in the EU, and health is 7.9% of US spending against 3.7% of EU spending. Same floor, a different bill in Austin and in Amsterdam. Second check: if your defend column lands at 20%, you've mis-sorted, because more than half of an average budget stays fixed under a shock.

Data: BLS Consumer Expenditure Survey 2024 and the Eurostat Household Budget Survey (reference year 2020). Two separate surveys with separate denominators, shown side by side rather than combined.

It's the same exercise as the fixed-cost audit that freed up over 4,000 Kč a month in our house, pointed at a different question. That Sunday asked which contracts were overpriced. This one asks which are load-bearing.

What Mine Looked Like When I Sorted It

The Golf is in the defend column, and not for car reasons.

I had every excuse to upgrade after our daughter arrived and didn't, so on paper it should sit in "would drop." But a reliable car that gets me to work and swallows a pram isn't the thing I'd stop paying for if my income disappeared. What lands in "would drop" is the replacement: the newer one, the taller one, the pro list I had four browser tabs open on. The car is floor. The upgrade is gap. Those two had been sharing one line of our budget for years.

The second house went the other way and it stung a bit. The spreadsheet calls it an asset. My calendar calls it a part-time job with a roof, and none of those hours has ever shown up in a yield calculation. It isn't floor, it isn't droppable either, and it's a commitment I chose years before I understood any of this.

Almost all the baby gear came second-hand, and the money was the smallest part of that — gear is a rounding error in a first year while childcare and housing are the lines that move. The unequal fun-money piles, my wife's deliberately bigger than mine, went straight into defend, which surprised me. They look like the definition of discretionary. They're actually what makes a savings rate around 50% survivable for years instead of eighteen months. A floor line that protects a marriage rather than a lifestyle.

And the finding that matters: our defended floor came out far smaller than our spending, and the gap between the two was almost entirely chosen, savoured and droppable. That's the whole reason a high savings rate has never felt like deprivation here. Not discipline. Just knowing which half is which.

Two Numbers Instead of One Date

What comes out of this isn't a tier. It's a pair.

One floor number, which is what you'd need if everything went sideways, and one chosen-life number, which is what you're aiming at. Then the gap between them, said out loud as something you're keeping on purpose rather than something you forgot to budget. Your target sits in that gap, and where it sits is a decision you can defend rather than a threshold someone published. Less tidy than a tier. Also yours.

Run both as scenarios rather than one date. The floor scenario tells you whether a plan survives being wrong, which, after one lucky stock pick and a crypto faceplant taught me I'm not a genius, is the only property of a plan I care about any more. It's also where the tracker I build earns its keep, so weigh it accordingly: the categories that produce the floor are the categories that feed the projection, so editing one recurring line shows the date move. The arithmetic itself is the easy part, and I've written it up in how to calculate your FIRE number.

Which is also my problem with percentage rules. 50/30/20 allocates income. It never asks which part of the 50 is permanent, and that's the only question that decides your number.

Now the Part Where You Argue With Me

The real objection isn't that I'm wrong about floors. It's that lean FIRE people are demonstrably adaptable, nobody treats their number as a vow, and flexible spending is the safety mechanism. The broken-promise framing is a straw man. It's a good objection, and the answer is in the opposition's own literature.

Michael Kitces measured what flexible withdrawal rules actually demanded of retirees in the bad sequences of US market history: real spending cuts of 28% for a 2007 retiree, 36% for 1999, 45% for 1936, 54% over a decade for 1965. Karsten Jeske found 59% for a 1966 retiree. Set those against the measured floor of $35,046, what the bottom fifth of US households actually spend in a year. A $40,000 lean household has $4,954 of headroom. Twelve percent. A 45% cut lands it at $22,000, below what the poorest fifth of Americans live on. A $120,000 household has 71% headroom, and the same cut lands at $66,000, still above the lean target.

Modelled on BLS Consumer Expenditure Survey 2024 spending levels (the lowest income quintile as the floor proxy), with historical cut sizes from Michael Kitces' Guyton-Klinger guardrails analysis.

Flexibility isn't a personality trait. It's arithmetic on the gap between your spending and your floor, and at a lean number that gap is a rounding error. The adaptability is real. There's just nothing left to adapt.

Two places where the objection beats me.

Human capital is the good one. Retire lean at 38 and a €15,000 shortfall has a dozen easy answers, and my arithmetic above ignores every one of them. Lean FIRE's flexibility is front-loaded: real at 38, gone at 68. Chetty and Szeidl cut the same way, incidentally — under a large shock, households do break their commitments, so the floor is stickier than a speed bump and softer than a wall. Both of those are timing arguments against me rather than refutations, and together they make a lean number a far better bet at 38 than it is a promise for the forty years after.

The second is that my method's output is unstable. In the same US Chase data, JPMorgan measured 63% of new retirees with more than 20% year-to-year spending volatility in their first three years, and 54% of 75-to-80-year-olds still living it. Pre-pandemic figures were similar, 55% and 51%, so it isn't a Covid artefact. A twelve-month sort has about a six-in-ten chance of being off by more than a fifth within three years, which is why the method has a third column and a rule about testing rather than guessing. A floor that's 20% wrong still beats a multiple from a page that can't agree with the next page on whether fat FIRE starts at $100,000 or $200,000.

Lean FIRE vs Fat FIRE Is the Wrong Argument. Price Your Floor.

Lean and fat are labels for where somebody's floor happened to land. The genre sells them as identities because an identity is easier to write a listicle about than a spending forecast is.

So here's the one thing I'd do this week if I were starting over. Sort twelve months of transactions into the three columns, then go and look at where your FIRE number came from. If it came from a tier, you built a target out of somebody else's floor.

Mine turned out smaller than I expected and the gap above it bigger, which is why I get to save half our income and still drive a car I like and take the holidays we actually enjoy. None of it was restraint.

Ask me again in four years, when the second house is closer to paid off and there's a school run in the calendar. The floor will have moved by then. Fine. It's a number I know how to re-measure, which is more than I could ever say for a label.