Forced Early Retirement Isn't What Scares Me. My Plan Quietly Assumes Twelve More Years of Earning.

The second house has about twelve years of payments left on it. I signed for it without once asking whether I'd be earning for twelve more years.

That's the whole article, and it took me an embarrassingly long time to see it.

I call myself coast-FIRE-by-40. It sounds like a claim about a portfolio. It isn't. It's a bet that I keep earning, at roughly this salary, in roughly this health, for about twelve more years, and forced early retirement is the name for what happens when that bet loses.

Everything downstream of that bet gets modelled to a precision the bet itself never gets. I've argued with myself about hedged versus unhedged. I've run the plan at 4% real and at 5% real and felt clever about the difference. I have opinions on fund domicile nobody asked for.

Number of times I have stress-tested the twelve years: zero.

Forced early retirement is real, but "earlier than planned" is not the same thing

Here's the number doing the rounds this year. EBRI's 2026 Retirement Confidence Survey, the 36th annual, 2,544 people, fielded in January: 46% of retirees say they retired earlier than planned, up from 40% a year earlier. Forty-eight percent stopped about when they meant to. Six percent stopped later.

Now the part the coverage keeps dropping.

EBRI asked the 495 people who left early why, and let them tick more than one box. Health problem or disability: 41%, up from 31% in a single year. Changes at the company: 35%, up from 22%. And sitting right in the middle of that list, the second-biggest answer of all: "I could afford to retire earlier than I planned," at 36%. Another 16% were offered a package and took it. Sixteen percent wanted to do something else.

EBRI's own headline is that 76% cited at least one reason outside their control, which is true, but because people could tick several boxes the 76% and the 36% overlap each other. You cannot read 46% as "46% got pushed out." Plenty of articles this year did exactly that, and I'd rather lose the scary version of a number than defend it.

Strip the drama and what's left is better anyway. For roughly half of people, the last day of earning was not the day they had written down. In either direction. Some got a diagnosis, some got a restructure, some got a windfall. The date is a distribution, not a decision.

That's the claim I'm actually making, and it's much harder to argue with than the one about victims.

Data: EBRI/Greenwald 2026 Retirement Confidence Survey, Fact Sheet 2.

Put the two sets of bars next to each other and the gap is the argument. Thirty-nine percent of workers expect to work to 70 or beyond, or never stop. Ten percent of retirees actually did. Twelve percent plan to stop before 60; 29% did. Median expected retirement age, 65. Median actual, 62.

Working longer is simultaneously the most popular retirement plan and the least frequently executed one.

The exit people imagine is not the exit they get

Allianz Life ran their own study this year, separate sample of 1,000, and they did something I wish more surveys did: they asked both groups the same question.

People still working, asked why they might leave early, said more time with family (36%), being financially ready sooner than expected (32%), and reducing stress (31%). A chosen, lifestyle-shaped exit.

People who had already left early, asked the same thing, said health issues that stopped them doing the job (30%) and unexpected job loss (21%).

Two populations, one questionnaire, completely different stories.

Data: EBRI/Greenwald 2026 Retirement Confidence Survey.

One more from Allianz, because it stuck with me. Thirty-five percent said that if they lost their job in the next six months, they'd probably just retire. Not job-hunt. Retire. Among boomers it was 58%. That isn't a plan either. That's a plan collapsing politely.

A working life is not a fixed length

I'm European, writing for people on both sides of the Atlantic, so I went looking for the mirror of the EBRI number over here. There isn't one. Europe has no equivalent survey, and the closest data is measured so differently that laying the two next to each other would be dishonest. Eurostat asks everyone who stopped working around their first pension for one main reason, and the overwhelming majority answer some version of "I reached the age." Different question, different denominator, not comparable.

The European evidence for the underlying point is better anyway, because it's about spread rather than sentiment.

Eurostat's expected duration of working life for 2025 is 37.5 years across the EU. That's the EU-wide figure, and the countries behind it disagree wildly: 44.0 years in the Netherlands, 32.7 in Romania. Eleven years of difference between two countries in the same union. Women in Sweden, 42.3 years. Women in Italy, 28.4.

Data: Eurostat, duration of working life 2025.

Then the figure that genuinely rearranged how I read my own date. Healthy life expectancy in the EU is 62.8 years for men and 63.3 for women, against total life expectancy of 78.9 and 84.1. The average European runs out of healthy years at roughly the age their state still expects them to be at work. That's not a policy complaint, it's a planning input, and it's why healthspan belongs in the financial plan and not in the wellness section.

Two more from the same Eurostat module, quickly. Across the EU, 19.4% of statutory pensioners took their pension early with a permanent reduction: a fifth of people paid a lifetime penalty to stop sooner. And 4.6% of everyone aged 50 to 74 draws a disability pension rather than an old-age one, rising to 11.5% in Estonia. Those people never appear in "retired earlier than planned" statistics, because they never got to retire. They got to stop.

Every guide to forced early retirement is written for the wrong reader

I read the whole first page of search results before writing this, which is a habit I'd recommend to nobody.

All of it addresses someone it has already happened to. Five things to do if you're forced into early retirement. Here's your game plan. Three critical moves. Not one page is written for the person who is a decade away with an untested assumption sitting in the middle of their plan.

And almost all of it is one country's plumbing: claiming ages, permanent reductions for filing early, health coverage gaps, severance and benefit rules. If you're not American you bounce in two paragraphs. If you are American, you get tactics for a situation you aren't in yet.

So here's my position. Most of what's published on this is aimed at the wrong reader at the wrong moment, and the single piece of advice it offers people who are still working is the least-executed plan in the whole dataset. Work longer. Thirty-nine percent expect to; 10% do. Seventy-four percent plan to work for pay in retirement; 31% managed it.

Nobody on that first page does the one calculation that's useful before the event, which is: what is a missing earning year actually worth?

So I ran it.

We argue about one percentage point of return and never model three missing years

Round numbers on purpose. Start with 200,000 invested. Add 1,500 a month. Assume 5% real return, monthly compounding, twelve years to go. Read it in euros or dollars, the arithmetic doesn't care.

Base case lands at about 653,400.

Now break it two ways.

Break it the way the internet argues: drop the return assumption by one percentage point to 4% real, keep all twelve years of contributions. You land at 596,500. That's 8.7% worse.

Break it the way nobody models: hold 5% real, but delete the last three earning years. You land at 595,100. That's 8.9% worse.

Data: Author calculation.

Those are the same number.

Losing a quarter of your remaining runway costs almost exactly what a one-point return haircut costs, and one of those two gets a hundred forum threads a week while the other gets set to zero variance and never mentioned again.

Delete five years instead of three and you're at 551,200, down 15.6%. Stop earning today and you're at 359,200, which is 45% below base. That last one isn't a trim, it's a different plan.

Be careful with what this is. It's arithmetic on inputs I picked, not a forecast, and the ratios move with your starting balance. If your pile is small relative to your contributions, the missing years hurt far more than mine do. If it's large, far less. That's the actual finding: the more of your plan still arrives in the form of future paychecks, the more your plan is a bet on your body and your employer rather than on the market.

This is a different animal from two things I've written before, and I'd rather draw the line myself than have someone draw it for me. Sequence of returns risk lives after the date: bad markets in the first years of drawdown. Income concentration is about the size of the position: one employer, one industry, most of your net worth. This one is about the length of the runway, and it's the only one of the three that has no name, no acronym and no chart in anybody's guide.

The reading I'd never taken

The check takes about a minute if you model your plan anywhere at all.

Take today's invested balance. Set future contributions to zero. Compound at a conservative real rate. Then look at two things: what that number supports at a 3.5% or 4% withdrawal rate, and how old you are when it gets there.

On my round-number example, the stop-today balance of 359,200 supports roughly 14,400 a year, or 1,200 a month, at 4%. Left alone at 5% real with nothing added, it takes about twelve more years to reach the base-case target. About fifteen at 4% real. So the honest answer isn't "I'm fine" and it isn't "I'm ruined." It's that if the earning stopped today the plan doesn't break, it slides roughly twelve years to the right. Which is a completely different sentence from the one I was carrying around, because I wasn't carrying one at all.

Then the second question, the one I'd genuinely never asked: at what age does the stop-today number become enough on its own? That age, not the date written in the plan, is the real end of your exposure. Everything between today and then is the stretch you're actually carrying.

Practically it's one setting change: contributions to zero, re-run the projection. I build a tracker for a living and it still took me longer to decide to look than to look. A spreadsheet does the same job. And if you've never worked out the denominator in the first place, go and do your FIRE number first, then come back.

Three commitments that quietly borrow from years I haven't earned

Plans don't get fragile because of the portfolio. They get fragile because of what you sign.

The second house. Roughly twelve more years of payments, priced off two salaries continuing on schedule. I already knew that building was nowhere near passive; it's a small business with a roof and it eats Saturdays in ones and twos. What I hadn't admitted is the other half of it. It's a twelve-year commitment underwritten by an assumption I never tested. We'd still buy it. I'd just have run the numbers with the earning stopped in year four before signing, and I didn't, because it never occurred to me that the earning was an input rather than a fact.

Upgrades that recur. A holiday is an expense. A bigger car on a monthly payment is a standing order wearing an upgrade's clothes. The first costs you once; the second signs your future self up for instalments out of income nobody has earned yet. Deliberate, savoured upgrades are the entire point of having money and I'm not arguing for a smaller life. I'm arguing that the recurring ones are a different instrument and deserve to be signed like one.

A savings rate that only clears at full health. Ours is high, and it assumes both of us working, neither of us ill, and nobody's parent needing daily care. That last one is the version almost nobody prices. Sixteen percent of EBRI's early retirees left to care for a spouse or family member, and 54% of caregiving workers say it already affects the hours they can work. It doesn't arrive as a diagnosis of your own. It arrives as a phone call about someone else, and it's the one scenario my wife and I have never put a number on.

The rule that falls out of all three is one sentence long. Don't sign a fixed commitment whose term outruns the horizon you'd actually bet money on. If I wouldn't put cash on being employed at full salary in year nine, I shouldn't sign something that needs year nine.

Design a plan that's finished at every point, not only at the end

A plan that only works if it's allowed to run to the last day isn't a plan. It's a wish with a spreadsheet attached. The repair isn't more money. It's four structural things, and all four are decisions you make once rather than habits you have to keep.

Keep fixed costs inside one income. Not total spending, which nobody can hold down forever anyway. Fixed costs: the contractual, cannot-cancel-this-quarter kind. Housing, insurance, loans, childcare, the subscriptions you'd have to phone a human to stop. If that total fits inside one of the household's incomes, an income shock is an inconvenience with paperwork. If it needs both incomes to clear, the same shock arrives as a crisis with a deadline attached, and you get to make your worst financial decisions in the worst month of your life. We are not all the way there, and the second house is exactly why. But it's the ratio I watch now, and I watch it more closely than the savings rate.

Hold the emergency fund as months, not as a round number. "Twenty thousand" is a vanity figure that stops meaning anything the moment your costs move. "Seven months of fixed costs" is a measurement, and it re-derives itself every time the fixed costs change instead of quietly going stale in a savings account.

Spend on staying employable, and budget it like an insurance premium. Small, annual, boring, not aspirational. Only one in ten older workers pushed out of a long-held job ever earns their old pay again, which makes being easy to re-hire the cheapest hedge on the entire menu, and considerably cheaper than anything a salesperson will offer you against the same risk.

Read the coast number as a floor, not a finish line. It isn't a milestone you cross once and tick off a list. It moves with your spending, so the useful version is a line you check you're still standing above, once a year, in the same sitting as everything else.

Now the part where you argue with me

Three objections. The first one is strong enough that I'll just concede it.

"The population trend is the exact opposite of your headline." It is. An ECB Economic Bulletin box published in March found that the effective age of leaving the labour market rose in almost every euro area country between 2022 and 2024, by about half a year on average, and that the rise was larger than the increase in statutory ages, so the law isn't what's pushing people. Workers aged 55 to 74 delivered 1.4 of the 1.7 percentage points of euro area employment growth since early 2022, and over 90% of that came from more people choosing to participate. Euro area employment at 60 to 64 is 53%, against Japan's 74%. Every aggregate says working lives are getting longer.

Fine. I'm not making the population argument. Averages rising is perfectly compatible with individual variance being wide. "People will stop earlier than they used to" is false. "Your own last day of earning has a spread you've never estimated, and your plan has it set to a single value" is a different sentence, and only the second one is about you.

"This is fear-selling." The best evidence in this entire piece is on that side, so here it is properly. EBRI's decumulation work, tracking households through the Health and Retirement Study to 2022, found that 21 to 22 years into retirement, 37% of the low-asset group, 48% of the middle and 42% of the high still held 80% or more of their starting money. Depending on the group, between 31% and 43% still held every cent of it. Asked why they hadn't spent it down, 37% said it simply seemed unnecessary. Meanwhile 73% of retirees say they're confident they have enough, and two in three say they're living the retirement they pictured. And the genre that publishes retirement fear is largely funded by people selling annuities, insurance and advice.

So if you finish this and respond by raising your savings rate and cancelling a holiday, I've made your life worse and I'd rather you hadn't read it. None of this is an argument to save more.

"Designing for it is the same as expecting it." No, and the difference is where the cost lands. The stop-today reading costs nothing in the good case: same contributions, same portfolio, same life. What changes is which commitments you sign. That's a constraint on the liability side of the balance sheet, not on the living side. You can hold this thought and still book the trip.

What changed, and what didn't

The standing order didn't change. The monthly purchase didn't change. The savings rate didn't change. If you came for a conversion story, there isn't one.

What changed is what the date means. It used to be a deadline. Now it's a range with a review attached, and two lines got added to the January session we already hold: read the stop-today number out loud, and check that nothing we've signed leans on year nine. Ten minutes, and neither line needs anyone to feel motivated, which is the only kind of money habit I trust.

I still like my job, and that's rather the point. Knowing the plan survives an early stop is exactly what lets me keep liking it, keep the monthly purchase boring, and spend money deliberately now instead of hoarding against a year that probably won't arrive. Freedom is having options, not having an exit.

Forced early retirement isn't really the risk. The risk is a plan that only works if it's allowed to run all the way to the end.

Mine was one of those until about a month ago.

The date in your plan is a forecast, not a decision. Give it an error bar before life gives it one.