Social Security: How It Works and When to Claim

Social Security is the largest asset most American retirees own, and almost nobody understands how the number is produced. This guide opens the formula, works the claiming math, and treats the solvency question honestly rather than as a talking point.

By Karanvir Singh, founder of Quantix Invest

What Social Security is and is not

Social Security is a government-run, inflation-adjusted, lifetime annuity funded by a payroll tax, with disability and survivor insurance attached. Understanding it as an annuity rather than a savings account clarifies almost every decision that follows.

It is not an account with your name on it. There is no pile of your money invested somewhere. The payroll taxes you pay today fund the benefits of today's retirees, and your future benefit will be funded by future workers plus whatever remains in the trust funds. The "trust fund" holds special-issue Treasury bonds, which are real obligations of the federal government but are not a private investment portfolio.

Three features make it unlike anything you can buy:

• It is indexed to inflation. Benefits receive an annual cost of living adjustment tied to a consumer price index. Commercial inflation-adjusted annuities barely exist in the US market and are expensive where they do.
• It lasts as long as you do. No sequence-of-returns risk, no longevity risk, no possibility of running out.
• It is progressive. The formula deliberately replaces a much larger share of a low earner's wages than a high earner's.

The funding mechanism: a 12.4% payroll tax for the retirement and survivors program, split evenly between employee and employer (self-employed people pay all of it), levied on wages up to an annual cap. That cap is $184,500 in 2026 and rises with average wages. Medicare's 2.9% is separate and uncapped.

Earning the benefit: credits and the 35 years

You qualify by earning credits. In 2026 one credit requires $1,860 of covered earnings, four credits is the annual maximum, and you need 40 credits (about 10 years of work) to be eligible for a retirement benefit. Credits determine eligibility only. They have no effect on the size of the benefit.

The size comes from your earnings history, and here is the rule that drives more planning decisions than any other: the formula uses your highest 35 years of indexed earnings. If you worked 30 years, five zeros are averaged in. If you worked 40 years, the lowest five drop out.

Everyone should pull their Social Security Statement from ssa.gov and verify the earnings record. Employers do occasionally misreport, name changes cause mismatches, and the correction window is generally about three years, three months, and fifteen days after the year in question, though clear evidence like a W-2 can reopen older years. An unrecorded year is a permanent reduction if nobody catches it.

Step by step: from wages to AIME

The calculation has three stages. Stage one produces your Average Indexed Monthly Earnings.

Step 1: index each year's earnings. Wages from 1990 are not comparable to wages from 2020, so each year is multiplied by an index factor that scales it to the national average wage level in the year you turn 60. Earnings after age 60 are used at face value, not indexed. This is why the formula is fair across generations: it measures your earnings relative to the economy of the time, not in nominal dollars.

Step 2: take the highest 35 indexed years. Sort, take the top 35, discard the rest, insert zeros if you have fewer than 35 years.

Step 3: divide by 420. Sum those 35 years and divide by 420 (35 years x 12 months). The result is your AIME, a monthly figure.

Worked example. Suppose a worker's 35 highest indexed years sum to $2,940,000. That is an average of $84,000 per year in wage-indexed terms, which is a solid middle-to-upper-middle career.

AIME = 2,940,000 / 420 = $7,000 per month.

Bend points and the PIA formula

Stage two converts AIME into your Primary Insurance Amount (PIA), the monthly benefit you would receive at exactly your full retirement age. The formula is a three-tier progressive schedule, and the thresholds where the rate changes are called bend points. They are re-set each year based on national average wage growth, and the ones that apply to you are locked in at age 62.

For someone turning 62 in 2026, the bend points are approximately $1,300 and $7,850. The formula:

Continuing the worked example for the worker with a $7,000 AIME:

• First tier: 90% x $1,300 = $1,170
• Second tier: 32% x ($7,000 minus $1,300) = 32% x $5,700 = $1,824
• Third tier: nothing, because AIME is below the second bend point.

PIA = 1,170 + 1,824 = $2,994 per month at full retirement age, or about $35,900 per year, indexed to inflation for life.

Notice the replacement rate: $2,994 monthly against $7,000 of monthly indexed earnings is 43%. Now run the same math for a low earner with a $2,000 AIME: 90% x 1,300 = 1,170, plus 32% x 700 = 224, for a PIA of $1,394, a replacement rate of 70%. And for a maximum earner with a $13,000 AIME: 1,170 + 2,096 + 15% x 5,150 = 1,170 + 2,096 + 773 = $4,039, a replacement rate of 31%.

Stage three applies the claiming-age adjustment, which is where most of the decision-making lives.

Claiming age: 62 vs 67 vs 70

Full retirement age (FRA) is 67 for anyone born in 1960 or later. It was 65 historically and was raised gradually by the 1983 amendments. Nothing magical happens at FRA except that it is the reference point where you receive exactly 100% of your PIA.

You may claim as early as 62 or as late as 70. Claiming later increases the monthly check permanently.

• Early claiming reduction: 5/9 of 1% per month for the first 36 months before FRA, then 5/12 of 1% per month beyond that.
• Delayed retirement credits: 8% per year (2/3 of 1% per month) for each year past FRA, stopping at 70. There is no benefit whatsoever to waiting past 70.

For an FRA of 67, the schedule against the $2,994 PIA from our example:

The spread from 62 to 70 is enormous: $3,713 versus $2,096, a 77% difference in monthly income for life, adjusted for inflation, guaranteed by the federal government. And because cost of living adjustments compound on the larger base, the dollar gap widens every year.

Despite that, claiming at 62 has historically been the single most common choice. Some of that reflects genuine need, ill health, or job loss in the early sixties, all legitimate. Much of it does not.

The break-even math, done properly

The naive comparison: claiming at 62 gives you an eight-year head start of $2,096 a month, so how long until 70 catches up?

Simple break-even, 62 versus 70. By age 70, the early claimer has collected 96 months x $2,096 = $201,200. From 70 onward, the delayed claimer receives $3,713 versus $2,096, an advantage of $1,617 per month.

Catch-up time = 201,200 / 1,617 = 124 months = 10.4 years. Break-even lands at about age 80.4.

Break-even, 62 versus 67. Head start is 60 months x $2,096 = $125,760. Monthly advantage from 67 onward is $898. Catch-up = 140 months = 11.7 years, so break-even at about age 78.7.

Break-even, 67 versus 70. Head start is 36 months x $2,994 = $107,780. Monthly advantage is $719. Catch-up = 150 months, so break-even at about age 82.5.

Now the refinements that the naive version omits.

1. Time value of money. The early claimer can invest the checks. If those dollars earn a real return of 2% per year, break-even against age 70 moves out roughly two years, to about 82. At a 4% real return it moves toward the mid-eighties. At 0% real, which is roughly what safe assets have delivered over many stretches, it stays near 80. The honest statement is that the answer is sensitive to an assumption nobody can pin down.

2. Life expectancy is not the average. The relevant number is not life expectancy at birth (about 79 in the US) but life expectancy conditional on reaching 62. Social Security's own period life tables put that at roughly age 82 for men and 85 for women. Roughly half of 62-year-olds will live past the break-even age, and a quarter of 65-year-old women will reach 90 or beyond.

3. The insurance framing is better than the break-even framing. This is the important point. Break-even analysis asks "which choice maximizes expected total dollars," which is the wrong question. The real risk in retirement is not dying early; it is living a long time and running out of money. Dying at 72 having claimed at 70 is not a financial problem, because you are dead. Living to 95 having claimed at 62 is a severe financial problem, because it happens to you.

4. The bridge strategy. A retiree who stops working at 65 does not have to choose between claiming early and having no income. They can spend down their portfolio from 65 to 70 to "buy" the higher benefit. This looks alarming (the portfolio shrinks) and is often the right move: you are converting volatile, taxable portfolio dollars into a guaranteed, inflation-adjusted, partially tax-free income stream at a very favorable implied rate. It also opens a window of low taxable income, ideal for Roth conversions.

5. When claiming early genuinely makes sense. Serious health conditions with a materially shortened life expectancy. No other income and no portfolio to bridge with. A much lower-earning spouse whose survivor benefit will come from the higher earner's record anyway (the lower earner claiming early costs the household little). Care responsibilities or involuntary job loss that make waiting impossible. These are real situations and there is nothing wrong with claiming at 62 in them.

6. The general rule for couples. Coordinate rather than deciding separately. The most common efficient pattern is for the higher earner to delay to 70 and the lower earner to claim earlier. The reason is survivor benefits: when one spouse dies, the survivor keeps the larger of the two benefits, and the smaller one stops. The higher earner's benefit is therefore effectively a joint-life benefit, and delaying it protects whichever spouse lives longer, which on average is the wife by several years.

Working while claiming: the earnings test

If you claim before FRA and continue to work, the retirement earnings test withholds part of your benefit. In 2026 the thresholds are approximately:

• Before the year you reach FRA: $1 withheld for every $2 of earnings above about $24,360 a year.
• In the year you reach FRA, for months before your birthday: $1 withheld for every $3 above about $64,800.
• From the month you reach FRA onward: no test at all. Earn anything.

The critical and widely misunderstood point: the withheld money is not lost. At FRA, Social Security recomputes your benefit upward to credit the months that were withheld. Over a normal lifespan you get most or all of it back. The earnings test is a deferral, not a tax, though it is often described as a tax and does discourage work.

Only wages and self-employment income count. Pensions, portfolio withdrawals, dividends, capital gains, interest, and annuity income are all invisible to the earnings test.

Spousal and survivor benefits

Spousal benefit. A spouse may receive up to 50% of the higher earner's PIA, if that exceeds their own benefit. Key details:

• The 50% is calculated on the worker's PIA at full retirement age, not on the worker's delayed amount. Delaying past FRA raises your own benefit but never the spousal benefit derived from it.
• The spousal benefit is reduced if the spouse claims before their own FRA, down to about 32.5% at 62.
• The spousal benefit earns no delayed retirement credits. There is no reason for a spouse to delay past their own FRA if they are collecting purely on a spousal basis.
• The worker must have filed for the spousal benefit to be payable.
• You cannot choose which to take. "Deemed filing" means claiming triggers both your own and any spousal benefit, and you receive the larger. The file-and-suspend and restricted-application strategies were closed for anyone born after January 1, 1954.

Divorced spouse benefit. If the marriage lasted at least 10 years, you are currently unmarried, and both of you are 62 or older, you may claim on an ex-spouse's record. If you have been divorced at least two years, the ex does not need to have filed. It does not reduce their benefit and they are never notified. Multiple qualifying ten-year marriages mean you take the best one.

Survivor benefit. On the death of a spouse, the survivor receives 100% of what the deceased was receiving (or entitled to receive, including delayed credits), if that is larger than their own benefit. The household then has one check instead of two, and the smaller one disappears. Survivor benefits can begin as early as 60 (50 if disabled), at a reduced rate, and unlike retirement benefits, a survivor can take one benefit first and switch to the other later, which creates genuine planning opportunities.

Also worth knowing: the Windfall Elimination Provision and Government Pension Offset, which historically cut benefits for people with pensions from non-covered government employment, were repealed by legislation enacted in January 2025. Affected retirees and their spouses now receive unreduced benefits.

How benefits are taxed

Up to 85% of your Social Security benefit can be subject to federal income tax, depending on "combined income," defined as adjusted gross income plus tax-exempt interest plus half of your Social Security benefit.

These thresholds are not indexed to inflation and have not moved since 1984 and 1993 respectively, so an ever-larger share of retirees is caught by them each year. This is a quiet, ongoing benefit cut that no one ever voted for a second time.

The planning consequence is the tax torpedo. Because each extra dollar of other income can make an additional 85 cents of benefit taxable, a retiree in the 22% bracket can face an effective marginal rate of 22% x 1.85 = 40.7% on portfolio withdrawals within the phase-in range. Roth withdrawals and, notably, tax-free HSA withdrawals do not enter combined income at all, which is why building Roth and HSA balances before retirement pays off precisely here. Municipal bond interest, despite being federally tax free, does count toward combined income.

A separate note: an additional standard deduction for taxpayers age 65 and over was enacted in 2025 and is scheduled to run through 2028, which reduces the tax bite for many retirees without changing the underlying thresholds. Rules like this expire, so check current law rather than relying on any guide, including this one.

Solvency, honestly

The claim that "Social Security will not be there" is the most common and most misleading thing said about the program. The accurate version is more specific and less dramatic.

The mechanics. Payroll taxes flow in and benefits flow out. For decades taxes exceeded benefits and the surplus accumulated in the trust funds, which now hold roughly $2.7 trillion in Treasury securities. Since 2021 the program has been paying out more than it takes in, so it draws on the trust fund. The Trustees' 2025 report projects the combined retirement and disability funds are depleted around 2034.

What depletion actually means. Not zero. Payroll taxes keep arriving from every working person. Those incoming taxes are projected to cover roughly 80% of scheduled benefits at that point, declining slowly to about 73% over the following decades. So the honest statement is: absent legislative action, benefits would face an across-the-board cut of about 20% in the mid-2030s. That is serious. It is not the program vanishing.

Why the shortfall exists. Demographics, mostly. Fertility fell, longevity rose, and the ratio of workers to beneficiaries has dropped from over 5 to 1 in 1960 to under 3 to 1 today and is heading toward 2.3 to 1. A pay-as-you-go system is a direct function of that ratio. A secondary cause is that rising wage inequality pushed more compensation above the taxable maximum than the 1983 reform anticipated.

What could fix it. The arithmetic is not exotic. The 75-year actuarial deficit is roughly 3.8% of taxable payroll, and each of the following closes a substantial share of it:

Any real reform will be a combination, and every option has losers, which is exactly why Congress has not acted. The 1983 reform, the last major one, arrived only months before the trust fund would have run dry, and it combined a payroll tax increase, a gradual rise in the retirement age, and the taxation of benefits. Political systems tend to act late rather than never.

Practical planning stance. Anyone currently over about 55 should plan on full scheduled benefits, since near-retirees are the group reform always protects. Someone under 45 might reasonably plan on 75% to 80% of the scheduled amount as a conservative base case, and treat anything above that as a pleasant surprise. Planning on zero is not conservatism, it is a forecast, and it is a forecast that would require a political failure with no precedent in the program's ninety-year history. Note that this is general education about a public program, not individualized financial advice; your own record, health, and household situation should drive the decision.

Bottom line and common mistakes

The three things that matter most: make sure your earnings record is complete and has 35 real years in it, think about claiming age as longevity insurance rather than as a bet, and coordinate as a household so the higher earner's benefit protects the survivor.

• Claiming at 62 by default, without running the numbers, because it is the first age available.
• Claiming early out of fear of insolvency, which converts a possible future cut into a certain permanent one.
• Never checking the earnings record and discovering a missing year after the correction window has closed.
• Deciding separately as a couple, ignoring that the survivor keeps only the larger benefit.
• Believing the earnings test is a permanent loss. It is a deferral and is credited back at full retirement age.
• Assuming Social Security will replace most of your income. It replaces roughly 70% for a low earner, 40% for a middle earner, and 30% or less for a high earner.
• Delaying past 70. Delayed retirement credits stop. There is no reason.
• Ignoring the tax torpedo when planning withdrawals, and paying a 40%-plus effective marginal rate that Roth or HSA dollars would have avoided.
• A spouse delaying past their own FRA for a purely spousal benefit that earns no delayed credits.