When Should You Turn On Social Security?

Ask ten people when to turn on Social Security and you’ll get two camps. One says take it at 62, because money in hand beats a promise. The other says wait until 70, because the monthly benefit gets so much bigger.

Both camps have math, and neither is wrong. They’re answering different questions. The wait-until-70 camp is asking how to collect the most dollars. The take-it-at-62 camp is asking what you’ll need, and when. The question that decides your retirement sits underneath both: what happens to the rest of your plan while you make this choice?

So this piece goes in order. First, how it works and what the usual chart leaves out. Then what the data shows when you run a whole retirement. Only then, what to consider. At the end there’s a calculator to run your own numbers.

Whenever you start, your benefit also gets a cost-of-living raise each January (3.6% for 2027). I assume 2.5% a year after that, so the income below keeps growing, just as yours will.

Waiting from 62 to 70 means giving up eight years of income, about $176,000 with the raises, to collect a bigger monthly benefit for life: by 70, about $3,626 a month instead of $2,047. The bigger benefit doesn’t pull ahead in total until about age 78.5. That’s the crossover most charts show.

But that chart assumes the early income sits in a drawer. They don’t. Either they’re invested, or they let your savings stay invested instead of being spent. Credit that money with 5% a year and the crossover for waiting to 70 moves from 78.5 to about 83.4. At 7%, about 87.7.

One popular way to say it: would you rather have the money at 62, or in your 80s? It’s a good question. It isn’t the whole question.

That table is conservative. It doesn’t build in future gains in longevity, and people who plan their retirements tend to outlive the averages. So the wait-until-70 camp has a real point: most people live past the simple crossover. Once the early income is invested, though, a single person’s odds fall to about even or worse.

But the odds cut both ways. Add up what each choice is expected to pay over a lifetime, counting each year’s income only by the chance you’re alive to receive it. Then bring every future dollar back to what it’s worth today, discounting each year by inflation plus 2% (roughly what safe, inflation-protected bonds have paid), because money that arrives sooner can be used or invested sooner. Measured that way, a single man’s expected lifetime total lands within about 3% anywhere from 62 to 70. For a single woman, waiting leans ahead by about 8%. For a married couple, the biggest edge, about 9%, comes from the higher earner waiting, because of the survivor’s income (the larger benefit a widow or widower keeps; more on it below).

That’s why smart people disagree. One widely cited study found more than 90% of people would come out ahead by waiting to 70, yet only about 10% do (Altig, Kotlikoff & Ye, NBER). Another found early claimers really are in worse health on average, though not by enough, on average, to erase what waiting pays (NBER). Both are right about Social Security by itself. Nobody retires on Social Security by itself.

Your savings pay for the wait. Retire at 62 and wait until 70, and eight years of living costs come out of your savings instead. Those withdrawals land in the first years of retirement, exactly when a bad market does the most lasting damage (that’s the subject of The Order of Things, another Forte Life piece). Here’s a married couple, both 62 and retiring now. The higher earner’s full-retirement-age benefit is $2,800 a month and the spouse’s is $1,400. They have $1,000,000 in their 401(k)s and need $70,000 a year to live on after taxes.

The household that waits draws almost twice as much from savings, and pays about twice the tax, for eight straight years. Only then does the bigger benefit let it draw less: at 70, with eight years of 2.5% raises, it needs about $20,000 a year from savings against about $47,000 for the couple that claimed at 62, and for a few years its tax nearly disappears. It doesn’t stay there. From 75 the IRS requires minimum withdrawals based on the size of the account, not on what the couple spends, and when one spouse dies the survivor files single, at higher rates. I ran this couple through the stock market’s actual returns 100 times, starting a retirement in each year from 1926 through 2025 (more on that below). By 70, the household that waited typically holds about $283,000 less in savings, in today’s dollars. In about 1 of every 8 of those starting years, its savings had fallen below $300,000 by age 70. The household that claimed at 62 never fell that low in any of them. That money is the cushion for everything a plan can’t predict: a roof, a health event, a bad decade.

Two more things the chart never shows:

• Taxes. Up to 85% of your benefit can count as taxable income, taxed at your normal rate, depending on all your other income. The income lines that trigger it ($32,000 and $44,000 for a married couple, $25,000 and $34,000 for a single filer) have never been raised for inflation. The $25,000 line set in 1984 would be about $80,700 today. So a bigger share of your benefit gets taxed almost every year. A couple with $40,000 of other income and $35,280 of benefits has about half the benefit counted as taxable today, about 69% in ten years, and 85% in twenty, at 2.5% inflation. Most states don’t tax Social Security, but most states with an income tax do tax the 401(k) withdrawals that pay for the wait. Tax on a 401(k) is never avoided, only timed. What you don’t take out now comes out later through the minimum withdrawals the IRS requires from 75, and whatever is left is taxed at your survivors’ rates: your spouse’s as a single filer, then your heirs’, usually within ten years. That’s why Roth conversions in low-income years, moving money into a tax-free Roth account at a low tax cost, can matter as much as the claiming age, and one reason the tax answer depends on the household. The full mechanics are in the Distribution Brief, another Forte Life piece.
• The survivor’s income. When one spouse dies, the household keeps the larger of the two benefits, not both. So the higher earner’s claiming age sets the survivor’s income for life. If the higher earner claimed early, the survivor still gets at least 82.5% of that person’s full benefit (CRS).

Congress has never let it happen. The closest call was 1983, and it acted only months before the money ran out. But look at what that fix contained: full retirement age raised from 65 to 67, a six-month delay in the cost-of-living raise, and, for the first time, income tax on benefits (SSA History). It was a benefit cut by another name. The next fix will likely look similar: not one dramatic cut, but trims through taxes, ages, and raises.

Either way, a cut takes the most dollars from the biggest benefit, and the household that waited spent eight years of its own savings to earn that bigger benefit.

I started a retirement in each of the 100 years from 1926 through 2025 and ran it on the stock market’s actual returns from there (a retirement that runs past 2025 picks up again at 1926). From each year’s market return I took out 2.5% for what it costs to be invested: 1.5% of activity drag, the return typically lost when money is actively managed and handled instead of simply left to track the index, plus a 1% advisory and fund fee. Both figures, and where they come from, are in The Actual Net Return, another Forte Life piece. Each of those 100 starting years ran through every lifespan Social Security’s table allows, with federal tax and a 5% state income tax figured each year and the minimum withdrawals the IRS requires from 75, then all over again with the 22% cut arriving in 2032. For each, I compared what the family had left at the end, after taxes and after subtracting any spending their savings couldn’t cover, and how often they ran short.

With full benefits, it’s close to a coin flip. With the cut current law calls for, claiming at 62 left the family more about two times in three.

Now the other half of the result: running short of money while still alive. Running short means savings were used up and Social Security plus any other income no longer covered what they needed to live on, in any year while either of them was alive. With full benefits, both claiming at 62 ran short in about 15% of the combinations, and the higher earner waiting cut that to about 7%. That’s the strongest case for waiting, and it’s real: a long-lived survivor living on the bigger benefit. That edge comes from long lives, and it doesn’t hold in every market. In the worst crashes right at retirement (1929, 1937, 1973), waiting ran out more often, because savings were being drained fastest just as the market fell. And with the 22% cut, the difference all but disappeared: about 21% either way.

Single retirees showed the same pattern. With a $2,400 benefit, $600,000 saved and $45,000 a year to live on after taxes, a single man claiming at 62 left more than waiting to 70 about 6 times in 10, and a single woman’s result was close to a coin flip. With the cut, claiming at 62 left more about 7 times in 10 for him and 2 times in 3 for her, and waiting no longer lowered the chance of running short.

Read it together. Waiting doesn’t remove risk. It moves it, out of your 90s and into the first eight years of retirement, when you’re selling the most savings and the market can do the most damage.

Two things tilt it back toward waiting, and they’re worth knowing. When I lowered the portfolio’s return by one point a year, roughly a more conservative mix, the higher earner waiting came out ahead about 60% of the time with full benefits, up from 50%. And Social Security’s life table is on the conservative side, so a longer-lived family would see waiting do somewhat better than these figures show.

The crossover chart answers how much. A retirement plan has to answer what if.

Here’s how I think about it with the families I work with. Income you’ve already collected is yours. It can stay in your plan and keep working, pay for the years you’re most able to enjoy, or pass to the people you love. The extra you wait for exists only if you’re alive to receive it and the rules still hold, and until then the government is holding that money, not you.

Some researchers call a delayed benefit the cheapest inflation-protected annuity you can buy (Center for Retirement Research), and it is priced well. The real question is whether you need to buy it, and what you sell to pay for it. When guaranteed lifetime income, like a pension or an annuity, already covers your basic expenses and protects against a long life, waiting buys that protection a second time, paid for with your own savings. In most of the plans I build, the numbers come out the same way: turn it on when you retire.

No rule fits everyone, but the data points in a clear direction for each of these.

• Lean toward claiming when you retire if your basic expenses are already covered by guaranteed income (a pension or an annuity), your health or family history points shorter than average, your savings will stay invested and keep working, you’d rather keep the money in your own control, or you’d rather collect what you can before a possible cut in 2032.
• Lean toward waiting if you’re the higher earner in a marriage and your spouse will depend on your benefit, you have no other lifetime income but enough savings to cover the waiting years without strain, you’re in excellent health with long-lived parents, or your savings earn bond-like returns.
• Still working before full retirement age? In 2026, Social Security holds back $1 of benefits for every $2 you earn above $24,480 (the earnings test) (SSA). It isn’t lost, but claiming early while working full time rarely makes sense. Past full retirement age and still working? There’s no earnings test at all.

“Please calculate my benefit on every record I’m eligible for: my own, my spouse’s, any former spouse’s, and any survivor benefit.”

• Bring the paperwork: marriage certificates, divorce decrees, and death certificates. Certified copies are usually available from the county clerk where each event happened.
• A former spouse counts if you were married at least 10 years and aren’t remarried. A late former spouse counts too, and remarrying after 60 doesn’t take that survivor benefit away (SSA).
• Told no years ago? Ask again. Records change, and people pass away without anyone notifying you.

Nate Butler

Forte Life · A Complete Financial System

nbutler@fortelife.net · 801-997-6861

Book a Retirement Income Diagnostic →

Scan to book