FIRE as a couple — and what happens if one of you dies
A couple's FIRE plan is one pot funding two lifespans, and the number that decides whether the survivor is all right — the survivor's pension — has no European rule behind it and no safe default. Seven wealthy countries have no permanent survivor's pension in their mandatory contributory system at all, and where one exists, two full careers typically cut it to about a quarter of the deceased's pension rather than a half.
So read the figure off your own provider's statement, and check how long it is paid: Czechia's runs for one year by default.
What a couple's plan is, and what most calculators do instead
Almost every "FIRE for couples" guide reduces two people to one person with a bigger income: add both salaries, add both savings, take the combined savings rate, multiply household spending by 25, done. None of that is wrong — a household really does have one pot and one spending figure, and pooling is the right first step.
It stops one step early. A couple is two working lives funding one household across two different lifespans, and three things follow that a single-person calculator has no way to express:
- Two retirement dates. "I stop at 55, you work to 62" is a common plan and a different plan from "we both stop at 58".
- Two life expectancies. The money has to last until the second death, not until the average one. If one of you is expected to reach 100 and the other 90, the horizon is 100. Lengthening the horizon usually raises the target rather than lowering it.
- A first death, decades before the plan ends. At that point one salary stops, that person's own pension stops, and the household's spending falls by far less than half.
One person does not need half of what two people needed
This is the part people underestimate, and it can be reasoned from a definition rather than a survey. The OECD's square-root equivalence scale holds that a household's needs grow with the square root of its size — so a single person needs 1 ÷ √2, about 71%, of a couple's income to reach the same standard of living. On the OECD-modified scale, which weights the first adult 1.0 and each further adult 0.5, it is 1 ÷ 1.5, about 67%.
square-root scale: 1 ÷ √2 ≈ 0.71 OECD-modified scale: 1 ÷ 1.5 ≈ 0.67
Rent or mortgage, insurance, heating, property tax, the standing charges and the car do not halve because one person is gone. So household income can fall by something close to half while household needs fall by roughly a third. The gap between those two is the whole subject of the rest of this guide.
There is no European survivor's pension
If you have read a page quoting a single "EU widow's rate", that number was invented by whoever wrote the page and attributed to the European Union. EU law decides which country pays and makes sure the payment follows you if you move. It does not decide how much, or for how long.
The European Commission is direct about it: national social security systems "remain different – the EU does not replace them with one single system", and every country is "free to decide who is to be insured under their legislation, which benefits are granted and under what conditions". Its own guidance for citizens adds the sentence most planning advice skips: "Not all EU countries pay survivors' pensions."
Two pieces of law set that boundary. Regulation (EC) No 883/2004 coordinates national systems rather than merging them — it exists, in its own words, "to respect the special characteristics of national social security legislation and to draw up only a system of coordination". And the equal-treatment directive, 79/7/EEC, does not reach survivors at all: Article 3(2) says it "shall not apply to the provisions concerning survivors' benefits".
Said precisely, because the imprecise version is wrong too. It is not that EU law ignores survivors — Regulation 883/2004 lists survivors' benefits inside its scope, which is exactly how a benefit earned in one member state can be paid to someone living in another. What does not exist is any European rule on amount, duration or eligibility. Those are national, and they differ more than most people expect.
The average is not your number
The figure you will meet most often is "about 50%". It is a real OECD figure and it is the wrong one for anyone reading this page. Here is the same sentence in full, from Pensions at a Glance 2025: survivor pensions pay "around half of the deceased's mandatory contributory pension to never-working survivors" — and "in the case of both partners with the same full career at the average wage, the survivor pensions replace about one-quarter".
Two careers halves it. Two careers is what a couple planning financial independence has.
The underlying modelling in the OECD's Pensions Outlook 2018 puts the same finding in numbers: the survivor replacement rate "decreases from 46% for a single-earner couple to only 22% when both individuals had the same baseline career". And it names the countries where the reduction goes all the way:
"In Austria, Estonia, Ireland, Japan, and the United States, own pensions fully eliminate entitlements to survivor benefits in the case of spouses having the same earnings history at the average wage." Not reduce. Eliminate. — OECD, Pensions Outlook 2018, ch. 7.
That is why Ember ships no default percentage anywhere in the product. For this audience the honest default is nearer nothing than nearer half, and a number on screen is an anchor whether or not it is labelled an estimate.
The ends of the range, and who is on them
The seven countries at the bottom are named in Pensions at a Glance 2025: "Coverage by permanent survivor pensions is included in mandatory contributory pensions in all countries except Australia, Latvia, the Netherlands, New Zealand, Norway, Sweden and the United Kingdom." At the top, the same report puts the figure at "more than four-fifths in Mexico, Poland and the United States".
Being on the bottom list does not mean a survivor there receives nothing. Residence-based basic pensions and safety nets still pay, and the OECD notes that in Canada, Denmark, Finland, New Zealand and the Netherlands those benefits help smooth consumption after a partner's death because the individual rate exceeds half the couple rate. What is absent is a benefit that exists because your partner died and keeps paying for the rest of your life.
These lists are not fixed. New Zealand abolished its survivor benefit outright in 2013; the UK (2010), France (2011) and the Netherlands (2015) each ended benefits for financially dependent spouses. A plan written today should not assume the rules of thirty years ago will still be there in thirty years.
Duration is the part nobody models
The headline percentage is the least of it. How long the benefit is paid is what decides whether a survivor's pension is a lifeline or a bridge — and the answer is frequently a bridge.
| Country | Default duration | What extends it |
|---|---|---|
| Czechia | 1 year | Caring for a dependent child or parent, third-degree disability, or reaching a set age |
| Germany (kleine Witwenrente) | 24 months | Converts to the große Witwenrente, which is not time-limited, on age, reduced earning capacity or a child under 18 |
| Sweden (omställningspension) | 12 months | Nothing — and it is only paid at all if you are under 67 when your partner dies |
Sources, in order: ČSSZ for Czechia — this project's home market; § 46 SGB VI for Germany, which is not subtle about it ("Der Anspruch besteht längstens für 24 Kalendermonate"); and Pensionsmyndigheten for Sweden. Germany's old unlimited rule survives only for marriages concluded before 1 January 2002 where one spouse was born before 2 January 1962 — § 242a SGB VI — which is to say, not for anyone planning early retirement now.
Model a one-year benefit as lifelong and you have overstated a survivor's income by an order of magnitude. That is why duration is a required field in Ember rather than an optional one.
And then the offset
Even where a survivor's pension is paid for life, most systems will not let you simply add it to your own. Czechia is a clean example: under the rule on concurrent pensions, the highest pension is paid in full and the others pay only half their percentage component. The general finding is the one above — the reduction is largest for a survivor with a full career of their own.
The United States: the higher of the two, not both
The US does something structurally different, and it catches people out. You do not receive your benefit and your spouse's. The Social Security Administration's own words: "If you're eligible for Survivor and another benefit, you'll choose the payment that's best for you. The payments won't be added together." You get the higher one. You can also switch later — start on the survivor benefit and move to your own at 70, when yours is largest.
That rule is unusually safe to model, because it needs no new number at all. Congress's own research service describes the mechanic as receiving "the retired-worker benefit plus a reduced auxiliary benefit amount equal to the full auxiliary benefit minus the retired-worker benefit, in essence receiving the higher" — which is arithmetic on two pensions you have already entered, not an assumption about either. Ember's "larger" setting is that identity, which is why it shows you no percentage: there is nothing to get wrong.
One caveat Ember does not model: timing. A survivor benefit claimed at 60 starts at 71.5% and reaches 100% only at survivor full retirement age, between 66 and 67, per the same SSA page. Ember applies the rule, not the claiming decision. If you plan to claim early, the figure you enter should already reflect that.
How to model it honestly
- Get the figure from a statement, not a percentage. Your own provider — or your partner's — can tell you what a survivor would actually be paid. No article can, including this one.
- Use the post-offset amount. The number that matters is what the household would receive after any reduction against the survivor's own pension, not the headline entitlement.
- Write down how long it runs, and model it ending on that date. If it is one year, it is one year.
- Check whether a workplace pension is single-life or joint-and-survivor. A single-life option stops on the day of death, whatever the state system does.
- If you cannot find a reliable figure, model none. That errs conservative, which is the direction a plan should err in.
And expect the rules to move. Poland now lets survivors combine pensions where it once forced a choice; Canada removed the benefit for separated couples who had split their entitlements; Japan is making its rules gender-neutral from 2028; Slovenia is raising both its rate and its eligibility age; Switzerland is expected to reform. All of that is from one OECD review of the last two years (Pensions at a Glance 2025, pp. 52–53). Re-read your own statement periodically; it is the only figure that is true about your household.
What Ember does, and what it does not
Ember models two people in the account app; the offline planner on this site models one person. The shape is two earners, one pot: two ages, two incomes, two dates for stopping work, two life expectancies, funding a single household from a single portfolio. The plan runs to the last survivor, and when the first of you dies that person's salary stops and so does their own pension.
For the survivor's benefit there are two settings, named for what they do rather than for a country. "The survivor keeps the larger of the two pensions" is the US rule stated exactly, and it renders no number at all. "A survivor's pension I'll enter" takes an amount you supply — the post-offset one from your statement — together with how long it is paid, which is required rather than optional. The default is not modelled, which never means "you will receive nothing".
No percentage, no per-country rate and no default amount appears anywhere in the product, for the reasons above.
What Ember does not model, stated plainly. Couples v1 does not model joint-versus-single tax filing — there is one flat tax blend, which cannot represent two filers, and that is exactly the effect US readers know as the widow's penalty. It does not apply career growth to a partner's income: the primary earner's career curve is not someone else's, so partner income is entered flat. It does not model separate portfolios. And it does not model claiming age for a US survivor benefit. The partner panel says all of this on screen, and the methodology page gives the formulas.
One more, because it is the kind of thing tools usually bury: if your device's clock is set to a US time zone, Ember may pre-select the "larger of the two" setting — and it tells you on screen that it guessed, and that it guessed from your browser. It reads only the time zone your browser already publishes to every website: no location lookup, no IP address, no third party, nothing stored or sent. It never changes a choice you have made.
Questions people actually ask
What happens to my pension when my husband dies?
It depends on which country insured him and which kind of pension it is, and the honest range is very wide. A workplace or private pension pays a survivor only if a joint-and-survivor option was chosen at retirement; a single-life option stops at death. A state pension is national law: the OECD reports that Australia, Latvia, the Netherlands, New Zealand, Norway, Sweden and the United Kingdom have no permanent survivor pension in their mandatory contributory system at all, while Mexico, Poland and the United States pay more than four-fifths of the deceased's pension. Read the figure off his own provider's statement rather than off any general article, including this one.
Is there a European survivor's pension?
No. EU law decides which country pays and makes sure a benefit follows you if you move, but it does not decide how much you get or for how long. The European Commission states that national systems "remain different – the EU does not replace them with one single system" and that each country is "free to decide who is to be insured under their legislation, which benefits are granted and under what conditions". Its own guidance for citizens adds the part most planning advice skips: not all EU countries pay survivors' pensions.
Do I get both my pension and my husband's when he dies?
Usually not in full — most systems make you choose or offset one against the other rather than stack them. In the United States, Social Security is explicit: "The payments won't be added together." You receive the higher of the two. The OECD found that in Austria, Estonia, Ireland, Japan and the United States a survivor's own pension fully eliminates the survivor benefit when both partners had the same earnings history at the average wage. That is the case which describes a couple who both worked full careers, which is most couples planning early retirement.
How long is a widow's or widower's pension paid?
Often for a fixed period rather than for life, and that is the detail most articles skip. Czechia pays the widow's or widower's pension for one year as standard, and beyond that only on specific conditions. Germany's kleine Witwenrente runs for at most 24 calendar months; the große Witwenrente is not time-limited but requires age, reduced earning capacity or a child under 18. Sweden's omställningspension is paid for twelve months and only if you are under 67 when your partner dies. Model a one-year benefit as lifelong and you have overstated a survivor's income by an order of magnitude.
How much does one person need after their partner dies?
Far more than half of what the two of you needed. On the OECD's square-root equivalence scale a household's needs grow with the square root of its size, so one person needs 1 ÷ √2, or about 71%, of a couple's income for the same standard of living; on the OECD-modified scale, which weights the first adult 1.0 and the second 0.5, it is 1 ÷ 1.5, or about 67%. Rent, mortgage, insurance, heating, property tax and the car do not halve because one person is gone. Household income can fall by close to half while household needs fall by a third.
Can one of us retire early while the other keeps working?
Yes, and it is the most common shape a couple's plan actually takes. Two working lives can end in different years, and each partner has their own retirement age and their own life expectancy — so the plan has to fund one household for as long as the longer-lived partner lives, not for as long as the person who ran the numbers lives. Modelling both people explicitly usually raises the target rather than lowering it, because the horizon gets longer and one salary stops before the other.
Run it on your own two lives
Add a partner, give each of you a retirement age and a life expectancy, and read the household year by year — including the years after the first death. Free, no sign-up needed.
Ember Pro — coming soon
Pro isn't on sale yet. Leave your address and I'll email you once when it opens — the planner stays free either way.
Stored by Ember only — never sold, never passed to an email service.