Wage indexing and the 35-year rule

If you have ever looked back at your work history and winced — the years you stayed home with kids, the stretch after a layoff, the months an illness took, the gig years you drove with no W-2 — this lesson is written straight at the worry underneath: “my benefit is already ruined, and there's nothing I can do about it.” It is one of the most common fears people carry about Social Security, and it is worth answering plainly, right at the top, before a single rule: the empty years in your record drag your benefit down, but they do not doom it — and the machinery you're about to learn is exactly why.

Here's the reassurance in one breath, and then we'll earn it: Social Security averages your highest 35 years, and a year you didn't work counts as a zero in that average. That sounds harsh, and the drag is real. But the same 35-year window means every future work year can overwrite a zero, one for one — and a zero that's really just *missing* wages (work you did that never got reported) is fixable. By the end of this lesson you'll be able to explain both machines the formula runs, and you'll have watched — to the dollar — what a zero costs and what overwriting one is worth.

Start with the problem the first machine exists to solve. Your benefit is built from a lifetime of earnings — for Ron, roughly forty years of them. But a dollar earned in 1995 and a dollar earned in 2026 are not the same size. Wages have roughly tripled over that span, so if the formula simply added up your raw paychecks, it would quietly punish everyone with a long career: your early years — earned when *everyone's* wages were smaller — would look tiny next to your recent ones, and drag your average down for no good reason.

Picture Ron's $30,000 paycheck from 1995. At the time, that was a genuinely good wage — solidly middle-class, the kind of year he was proud of. But laid next to a 2026 salary, the raw number looks small, almost like a bad year. That's the distortion, and it would fall hardest on older workers and on anyone whose best earning years were long ago. So before it averages anything, Social Security runs a fairness step called wage indexing: it restates each past year's earnings in today's wage terms, so a good year in 1995 is remembered as the good year it was. That's the whole purpose of the first machine — it makes decades comparable.

So how does the restating actually work? Each past year gets multiplied by its own indexing factor — a number that scales that year's wages up to current levels. The yardstick behind every factor is the AWI, the national average wage index: SSA's official measure of what the average American earned in a given year. The most recent one, the 2024 AWI, is $69,846.57 — about 4.84% above 2023's $66,621.80. Your old wages are measured against that yardstick.

The factor for any earlier year is a simple ratio: the AWI of the year you turn 60, divided by the AWI of that earlier year. Take Ron's $30,000 from 1995. Multiply by an indexing factor — using a round, illustrative factor of 2.1 — and that year is restated as $63,000 in today's wage terms. That $63,000 is now what the formula “sees” for 1995. To be clear, 2.1 is a teaching stand-in: your real factor for any year comes straight from the AWI series and *your own* age-60 year, and it will differ. SSA's own published example makes the mechanic concrete: for a worker turning 62 in 2024, $20,000 earned in 1990 indexes to $60,676.42 (a real factor of about 3.034).

Why the year you turn 60? Because that's the anchor the whole scale is pinned to — technically, two years before the earliest age you could claim retirement (62). Every earlier year is lifted up to the wage level of your age-60 year. And notice the last two rows of the table, because they hold a surprise worth its own beat.

Here is the part that catches almost everyone off guard, in a good way. Earnings in the year you turn 60 and every year after are counted at face value — their indexing factor is exactly 1. The year you turn 60 sets the yardstick (it divides by itself), and from 61 onward your wages are never scaled down to match some earlier era. A paycheck you earn at 61, 63, or 65 enters the formula at its full, actual dollar amount.

Sit with why that's kind. The years right before and after retirement are, for many people, their highest-earning years — and the formula lets those years count at their full modern size, not shrunk. For Ron, still working at 63 and weighing when to claim, this means his recent paychecks land in the average at full weight. It also quietly rewards working a little later if that's what your life calls for — though whether to do that is a personal decision we never push (it's weighed honestly in Lesson 145). The mechanic here is simply: old years get lifted up; recent years are already at today's size, so they're left as they are.

Now the second machine, and it's the one the fear at the top of the lesson really lives in. Once every year has been indexed, Social Security ranks all your indexed years and keeps only the top 35. That's the 35-year rule. Not your *last* 35 years, not the 35 you'd pick — your highest 35 indexed years, whichever ones they are. Everything else is set aside. Then those 35 are added up and averaged (the ÷ 420 months step, worked in full in Lesson 24).

That single rule cuts two very different ways depending on how many years you've worked, which is why we follow two people. If you have more than 35 years — like Ron — the rule is a gift: your lowest years simply fall away and don't count. If you have fewer than 35 years, the empty slots don't just disappear; they get filled with zeros, and a zero is exactly what it sounds like — a year of $0 dragged into your 35-year average. The rest of this lesson is about what that costs, and what you can do about it. First, Ron's happy case.

Ron worked about 40 covered years, steadily, at a solid wage. When the selection machine ranks those years and keeps his highest 35, his five lowest years fall away entirely — and here's the quiet magic: because he has more than 35 real years, not one zero ever makes it into his average. Every one of his 35 kept slots holds a real, indexed year. His early, lower-paid years don't hurt him; they're simply the ones that got shed.

That's why Ron's benchmark is solid. His 35 highest indexed years sum to $2,730,000 — the locked figure we carry through the course. Averaged over 420 months, that's an AIME (Average Indexed Monthly Earnings) of $6,500 (the full averaging is Lesson 24). Run through the benefit formula (that's Lesson 25's job — we only borrow the result), his PIA — Primary Insurance Amount, the benefit at Full Retirement Age — is $2,825.80 a month. No zeros, no drag, a clean 35-for-35 record. Hold that $2,825.80: it's the no-zero yardstick we'll measure the cost of zeros against.

Now the other side. Tasha Nguyen is 31, and she's been driving rideshare and delivery for a handful of years. If SSA computed a retirement benefit for her today, only about 10 of her 35 slots would hold real earnings — the other 25 would be zeros. On paper that looks devastating. It is not, and the reason is a distinction worth making sharp: not all zeros are the same thing.

1. Unlived zeros — years you simply haven't worked yet (or didn't work). Most of Tasha's 25 zeros are this kind: she's 31, and the years that will fill those slots are still ahead of her. These aren't a verdict on her record; they're placeholders, and each future covered year overwrites one. A genuine gap year — caregiving, a layoff, illness — is also this kind: permanent only if nothing later replaces it.
2. Fixable zeros — work you actually did that never got counted. A year can show $0 not because you didn't work, but because wages were never reported, or misreported. Tasha has already lived this: a platform misreported her 2024 earnings, which she caught and corrected (that's the arc of Lesson 17). A zero like that is an error, not a fact — and you fix it with your own proof (W-2s, pay stubs, tax returns), for free.

So the honest picture for Tasha is not “ruined.” It's “mostly unwritten” — a window with a lot of open slots and a lot of working years still to come. The only way to tell a fixable zero from an unlived one is to look at your record (that's Lesson 16), because the two demand completely different responses: one you *correct*, the other you *fill over time*. Neither is a life sentence. But before the reassurance goes any further, let's be unflinching about what an unfixed zero actually costs — because vague comfort helps no one.

To measure the pure cost of a zero, we need to hold everything *else* constant — same earning power, only the empty slots differ. So here is a labeled illustrative stand-in: a worker shaped like Ron's earning power, where each filled year carries $78,000 of indexed earnings (that's exactly Ron's per-year level — his $2,730,000 spread over 35). This is not Ron. It's a teaching construction that lets us knock out five slots and watch *only* what the zeros do.

Give this stand-in 30 filled years and 5 zeros. The indexed earnings sum to $2,340,000 (that's 30 × $78,000; the 5 zeros add nothing). Averaged over the same 420 months, the AIME is $5,571 — and note the rounding, because it's real law: 2,340,000 ÷ 420 is $5,571.43, and SSA floors it to the next lower dollar, $5,571. Run that through the benefit formula and the PIA is $2,528.60 a month.

Now the comparison that matters. Ron's no-zero benchmark is $2,825.80; the five-zero stand-in lands at $2,528.60. The difference — $297.20 a month — is the entire price of those five zeros, and because the earning power was held identical, none of that gap is about lower pay — it is purely the cost of the empty slots. Over a year that's about $3,566, every year, for life. That is a real number, and it deserves to be said without sugar-coating: zeros drag, and the drag is not trivial. The next section is the other half of the truth — the part the fear leaves out.

Here's the disarming half. Take that same five-zero stand-in and add one more indexed year — say the $63,000 year from our indexing example — dropped into one of the empty slots. The sum climbs from $2,340,000 to $2,403,000. The AIME goes from $5,571 to $5,721 (2,403,000 ÷ 420 = $5,721.43, floored). And the PIA rises from $2,528.60 to $2,576.60 — a gain of $48.00 a month, for life, from a single year that overwrote a single zero.

That's the whole answer to the opening fear, made concrete: an empty slot is not a closed door. Because the window is a rolling highest-35, any covered year you earn from here can climb into it and shove a zero out. For someone with zeros to spare, the arithmetic is friendly — in fact, replacing *any* single zero with that $63,000 year adds the same +$48.00, because one extra year lifts the AIME by a fixed $150 and the formula pays 32 cents on that dollar in this range (why it's 32% is Lesson 25's bend-point story). Zeros drag — but they're the most repairable part of your whole benefit.

Two rounding conventions ran silently through the math above, and they're worth naming because they're actual SSA law, not estimates. First, the AIME floors to the next lower dollar. Our stand-in's average was truly $5,571.43, but SSA drops the cents and uses $5,571 — and it never rounds up. Second, the PIA rounds down to the next lower dime.

Here the dime rule shows its teeth on the locked benchmark: Ron's formula output is actually $2,825.88, and SSA floors it to $2,825.80 — those 8 lost cents are the rounding biting. The illustrative PIAs happened to land already dime-exact ($2,528.60 and $2,576.60), so nothing was lost there — but that's luck, not a rule. (A later step you'll meet at Lesson 25 rounds the *payable* check down to the next lower dollar too.) Small as they are, these floors are why a hand-computed benefit that ignores them will always run a few cents high.

All of this points to one action, and it's the practical heart of the lesson: look at your earnings record now, not at claiming time. Every past year sits there, and the zeros in it fall into the two kinds we met — the fixable (missing or misreported wages) and the unlived (years still ahead, or genuine gaps). You want to catch the fixable ones early, because correcting a wage error gets harder as the paperwork ages — old employers vanish, records get thin.

• To read your record — the year-by-year table of your covered wages, and where a suspicious zero hides — is Lesson 16. Your personal copy lives in your my Social Security account and on your Statement.
• To fix a wrong year — the correction process and the proof SSA accepts (W-2s, pay stubs, tax returns; the SSA-7050 form) — is Lesson 17. It's free; anyone charging you to “fix” your record is running a scam (next section).
• To fill unlived zeros — that's just future covered work, at whatever pace your life allows, each year overwriting a zero. And the machinery that keeps re-running your benefit automatically as new years come in — even after you've claimed — is Lesson 28.

One reassurance on scope before we close the teaching: it's one federal formula, identical in all 50 states. Indexing, the AWI, the highest-35 rule, the rounding — none of it changes when you cross a state line. Your record is your record, wherever you live. What we've *not* done — and won't — is compute *your* benefit; the honest source for that is your own Statement, and for your own situation a free call to the SSA at 1-800-772-1213.

The moment people learn that zeros drag their benefit, a very specific scam comes looking for them — and it preys on exactly the fear this lesson set out to calm. The pitch: “Pay us to optimize your 35 years,” or “We can remove the bad years from your record,” or “There's an error dragging your benefit — just give us your my Social Security login and we'll fix it.” Every version is false, and the last one is identity theft with a helpful voice.

Here's the tell that collapses all of them: the only way to change your record is SSA's own free process, using your own proof — and no one can “remove” a real year or a real zero. A year you didn't work can only be overwritten by real future earnings, never deleted by a service. And SSA will never charge you to “optimize,” “clean up,” or “recalculate” your record, and never needs your password or a verification code to do it. So anyone selling a shortcut to “fix your years” isn't fixing anything — they're after your login and your SSN. Don't pay, don't share, and start any real correction yourself (Lesson 17).

If some of this landed on a tender spot — a stretch you spent raising kids, a layoff you didn't choose, cash work that never got reported, a year illness took — this beat is for you, and it's distinct from the scam warning above. That was about a fraudster; this is about your own footing. Almost no one arrives at 35 clean years in a straight line. A record with gaps is the normal shape of a real life, not evidence you did anything wrong.

Say the load-bearing line plainly: a zero is a placeholder, not a verdict. It marks a slot that's still open — and an open slot can always be filled, at whatever pace your life allows, each future year overwriting one. A zero that's really missing wages is fixable with your own proof (Lesson 17). And you never owe anyone an explanation for a gap year — the formula doesn't ask why a year is empty, and neither does anyone who matters. Whether you fill more slots by working longer is entirely your call — this lesson shows the mechanics and never pushes the decision (that's Lesson 145, laid out without steering).

The questions this topic raises most often, answered plainly — each one echoes a beat above, and several get a fuller home later in the course.

I stayed home with my kids for six years — is my benefit ruined?

No. Those years may show as zeros in your 35-year average, and zeros do drag — but they're placeholders, not a verdict. Because the formula keeps your highest 35 years, any future covered year can overwrite a zero, one for one. And if any of those “zeros” is actually work that went unreported, that's an error you can fix with your own proof (Lesson 17). Ruined is the wrong word; unwritten is the right one.

Is it my best 35 years or my last 35 years?

Your highest 35 indexed years, ranked across your whole working life — not your last 35 and not your most recent. A strong year at 24 can outrank a lean year at 58. This is good news: an early good year is never “too old to count,” and a low-earning late year can't pull your average down unless it actually cracks your top 35.

I've worked fewer than 35 years total — what happens to the empty slots?

They're filled with zeros, and the average is still divided over the full 35 computation years (420 months) — the calculation doesn't shrink to match a short career. That's the drag. The fix is simply time: each additional covered year replaces a zero and lifts your average, until you've filled all 35. On our illustrative stand-in, one $63,000 year overwriting a zero was worth +$48.00 a month for life.

Do my teenage or early-career jobs even count?

Yes — if the work was covered and reported. And this is exactly what indexing is for: it lifts those small old paychecks into today's wage terms so they're compared fairly. A $30,000 year from 1995 doesn't sit in your average as a tiny number; it's restated (illustratively, ×2.1) as about $63,000. Whether an early year makes your final top 35 depends on your other years — but it absolutely counts, and indexing gives it a fair shot.

Are my old wages counted at their old, smaller value?

No — that's the whole point of the first machine. Earnings before the year you turn 60 are indexed up to current wage levels using the AWI (the 2024 AWI is $69,846.57). And here's the pleasant surprise: earnings in the year you turn 61 and later are counted at face value — never scaled down. So old years get lifted, recent years are already current-sized and left alone. Nothing is stuck at its 1995 size.

Does working one more year always raise my check?

Only if that year beats a year already in your top 35 — and a zero is the easiest thing to beat, so if you have empty slots, yes. But if your 35 are already full (like Ron's) and your new year comes in below your current lowest-of-35, it doesn't enter the average and your check doesn't change. “One more year” overwrites your lowest slot; it isn't a guarantee. Whether to work longer is a personal decision the course never steers (Lesson 145).

Can I pay a service to remove my bad or zero years?

No — and anyone who offers to is running a scam. No one can delete a real year or a real zero from your record; the formula only ever counts real reported wages and real SSA corrections. A zero from a year you didn't work can only be overwritten by real future earnings. Correcting a genuine wage error is free and done by SSA with your own proof (Lesson 17). If someone wants a fee — or your login — to “optimize your 35 years,” they're after your identity, not your zeros.

One idea is worth feeling with your own hands before you go: how zeros move the check, and how filling one moves it back. The lab below builds the illustrative 35-slot average from this lesson — each filled slot at $78,000, shaped like Ron's earning power. Drag the zeros slider and watch the AIME and PIA fall; press replace one zero with a $63,000 year and watch the check climb. It's a stand-in to *show the machine*, not your own record and not a prediction.

Watch the two anchors as you play. Slide to 0 zeros and the benefit lands on $2,825.80 — exactly Ron's locked no-zero benchmark, the same 35-for-35 record from earlier. Slide to 5 zeros and it drops to $2,528.60 (the $297.20 drag). Replace one and it rises $48.00 to $2,576.60. Notice the honest mirror the lab enforces: at 0 zeros there's nothing left to replace — once the 35 are full, more years only help if they beat your lowest kept year. When any of this touches your real life, the next step isn't a stranger's calculator; it's your own Statement (Lesson 16) and, if you want a human, the SSA at 1-800-772-1213.

Every term introduced or leaned on above, in one plain line each — the vocabulary you can carry into Lessons 24 and 25.