One Field, Thirty-Five Years: Put the SSA Earnings Record in the Plan, Not a Typed Estimate

The discovery meeting is going well. You reach the Social Security line of the fact-finder, and your client — 52, a school aide, in the country since she was nineteen — reads you a number from a statement she printed two years ago: $2,340 a month. You type it into the one field your planning software gives you, the field labeled *estimated benefit*, and the projection updates. The plan now rests on that number for the next forty years of modeled life.

Here is what that field just accepted without asking a single question: whether the record behind the number is complete, whether the estimate's own assumptions match her actual plans, and whether the wages she earned in her twenties were ever credited to her at all. A typed benefit estimate is not a data point. It is a claim about thirty-five years of earnings history — and the software believed it on your say-so.

A Typed Number Is a Claim About 35 Years

Social Security computes the retirement benefit from the worker's highest 35 years of indexed earnings. Sum them, divide by 420 months, and the result — the average indexed monthly earnings — runs through the bend-point formula to produce the primary insurance amount. Two properties of that computation matter for the way you capture the input:

• Years with no credited earnings enter the average as $0. They are not skipped. A worker with 26 earning years and 9 zeros has those 9 zeros sitting inside the 420-month window, pulling the average down for life.
• The statement's estimate assumes the future. The printed number generally assumes the client keeps working until claiming age at roughly her recent earnings level. For the client who intends to stop at 58, or drop to half-time to look after a parent, the printed estimate quietly overstates the benefit — because the assumed future years were going to replace some of the zeros, and now they won't.

The single benefit field in most planning software can represent none of this. It cannot distinguish a verified number from a guess, a complete record from one with nine holes, or an estimate whose assumptions match the plan from one that contradicts it. Whatever you type is what every downstream page — income floor, claiming analysis, survivor projection — inherits as truth.

Where the Zeros Come From in the Households You Serve

Every book has some zero years in it. In first-generation bilingual households they cluster, and they cluster for reasons a standard fact-finder never surfaces:

• Caregiving years. The years spent raising children or looking after parents post as zeros, and in the families you serve that work often ran a decade or more — frequently for the spouse whose benefit you are about to type in from memory.
• Years worked in another country. Earnings from Mexico or elsewhere before immigration do not appear on the U.S. record. A client who arrived at 30 has a structural block of zeros no future raise will touch — only additional U.S. earning years can replace them.
• Missing W-2s. Employers that folded, paid in cash, or misreported. The client remembers the job; the record shows a zero or a fraction of the real wage.
• Name and number mismatches. This one is nearly invisible and disproportionately hits your client base. Wages reported under a maiden name after a marriage, a hyphenated surname the payroll system truncated, both surnames in the Mexican convention where the employer filed only one, a transposed digit in the SSN — when the name and number on a W-2 don't match SSA's files, the wages are not credited to anyone. They sit in SSA's earnings suspense file, which has accumulated hundreds of billions of dollars in uncredited wages over the decades. On the client's record, a year she worked full-time reads as $0.

The client cannot report any of this to you, because she has never read the year-by-year table. She read you the big number on page one. The zeros live on page two.

The Record Is a Discovery Document

The fix is procedural, and it is cheap: treat the SSA earnings record — not the benefit estimate — as the discovery document. The estimate is an output; the record is the data.

• Make the record pull a standing intake item. Every client with a claiming decision ahead creates a *my Social Security* account at ssa.gov/myaccount and pulls the full earnings record — the year-by-year table, not just the estimate page. For clients who avoid government websites (many in this demographic have learned caution the hard way), walking through the account creation together in the first meeting is ten minutes that pays for itself for a decade.
• Read the 35-year table with the client, in the language the household actually reads numbers in. Go year by year. "2003 shows zero — where were you working?" is the single highest-yield question in the review. It surfaces the cash-paid restaurant years, the maiden-name years, the employer that never filed. A zero the client can explain is a plan input; a zero she cannot explain is a correction case.
• Flag the correctable years now, not at claiming. SSA's general time limit for correcting the record is 3 years, 3 months, and 15 days after the year in question — but the exceptions are broad, and a correction supported by evidence (a W-2, pay stubs, a tax return) can be made far later. The evidence is the constraint: a shoebox of pay stubs from 2004 is findable today and gone in fifteen years. The claiming meeting is decades too late to start that search.
• Enter the record's reality into the plan. Count the zeros inside the top-35 window. Then model the futures the record actually offers: the benefit if she stops at 58 with the zeros intact, the benefit if three more full-time years replace three zeros, the benefit if the two maiden-name years get corrected. Replacing a single zero with a $40,000 year moves the 35-year average by roughly $95 a month, and the formula converts a meaningful slice of that into a raise that lasts for life — the survivor's life included, since the survivor keeps the larger check.

What This Changes in the Software

Once the record is the source, the number in the software's benefit field changes character. It stops being "whatever the client remembered" and becomes a verified figure with provenance — this estimate, from this record, pulled on this date, with these zeros counted and these assumptions checked against the client's actual stop-work plan. When the projection shows the claiming analysis, you can defend every dollar of the input, in either language, to any member of the family.

It also changes what you model. A record with four correctable years is not one scenario; it is two — as-is and repaired — and the delta between them is a concrete, dollar-denominated agenda for the next twelve months. That is planning work no typed estimate can generate, because the typed estimate erased the very information the work comes from.

This is the same standard WiseNest holds families to on the consumer side: before a benefit number enters a plan, the app sends the user to ssa.gov/myaccount to get the real one. For the advisor, WiseNest Connect picks up from there — the verified benefit lands in one household model where both spouses' checks, the claiming ages, and the survivor path are computed together and presented in English and Spanish side by side, so the abuela who earned those thirty-five years can read the result of the record she built.

Monday Morning

• List every client within fifteen years of claiming whose plan holds a Social Security number you did not verify against the record. That list is the exposure.
• Add the earnings-record pull to intake and to every annual review north of 50. The record changes every year; the plan should notice.
• Put the year-by-year read-through in the next review for your first-generation households. Budget twenty minutes. Ask about every zero.
• Share the consumer zero-years article with the families it fits. A client who has already counted her own zeros arrives at the claiming conversation as a partner, not an audience.

The claiming conversation is the most consequential single decision in most of your clients' plans. It deserves better than a number somebody typed from memory into a field that never asked where it came from.