Population-Weighted Results

Companion to the headline study. Same corrected engine (August 2026), same cells — weighted to the U.S. full-time earnings distribution, July 2026.

Why this layer must be read differently

A previous population study was retracted. It could ignore its own weights, because every scenario won — so any weighting returned 100%. That is no longer true. On the corrected engine, who wins depends on income and, overwhelmingly, on investment behavior. So this analysis follows two disciplines:

1. Results are reported separately for each investment-behavior band, never blended. A single "X% of Americans" number would require asserting how workers actually invest, and that assertion would carry the whole answer. We refuse to make it invisibly.
2. Every weight-sensitive number is stress-tested across income-distribution assumptions, and flagged when it moves.

Method, in short

Each worker's peak career income is drawn from a log-normal distribution calibrated to the Census full-time year-round median ($63,360) and to the SSA's anchor that about 8% of covered workers earn at or above the taxable maximum (fitted σ = 0.73). Incomes are capped at the taxable max ($176.1K) and floored at the model's $20K minimum. Income is the only weighted dimension, because the corrected engine is cohort-invariant and demographics move the ratio only about ±10%. The 1980 cohort is the anchor. (Full parameters at the foot of this page.)

The results, by band (July 2026)

*Yellow flag = DC wins on present value, but the 10th-percentile-market monthly income falls below the Social Security benefit. †Weight-sensitive — see below. ‡Uses the glidepath's dispersion as a proxy.

Reading each band honestly

Treasuries — Social Security won for everyone. No income level cleared 1.0× deterministically, the weighted Monte Carlo win rate was half a percent, and the median retiree would draw 42% of the Social Security income. There was no population for whom an all-cash DC system was a good trade.

Bonds — a 19% deterministic "win" that evaporated on inspection. Workers above roughly $119K peak income beat SS on the single historical path. But three qualifiers gut that number: it was the most weight-sensitive figure in the report, ranging 10% to 28% across reasonable income-distribution assumptions; those same cells won only about 6% of resampled market histories; and every one of those deterministic winners was yellow-flagged. The stable, honest bonds figure is the Monte Carlo one: about 3% of population-weighted market outcomes beat Social Security. A bonds-only DC system loses.

Equities-Conservative — won the war, risked the monthly paycheck. Every worker beat SS on present value in effectively every simulated history. But an estimated 78% of workers carried a yellow flag — in an unlucky market decile, their drawable income fell below what Social Security would have guaranteed, even though lifetime value was higher. The lowest-income workers, for whom the guarantee matters most, were the most exposed. A conservative equity posture bought lifetime value but not income security.

The glidepath — the realistic base case, and a clean sweep. 100% of workers won deterministically; 99.9% population-weighted Monte Carlo; a median inheritable balance of $1.5 million in today's dollars against Social Security's $0; a median income of 3.7× the SS benefit; and zero yellow flags — even an unlucky-market retiree out-earned Social Security at every income level, including the floor. The standard target-date structure was the posture that won on value and on downside income for essentially every worker.

Why the weights barely mattered — except where they did

For the glidepath and equity bands the conclusion was invariant (every measured income won nearly every trial, stress range 99.8–100%). For the fixed-income bands the deterministic share was weight-sensitive (bonds 10–28%), which is exactly why those bands are led by the Monte Carlo figure, stable to within ±0.4 percentage points across every stress test.

The two-sentence version

Weighted to the U.S. full-time earnings distribution, a worker on a standard target-date glidepath beat Social Security on lifetime value in about 99.9% of simulated market histories, retired on roughly 3.7× the monthly income, out-earned the guaranteed benefit even in an unlucky market, and left a median seven-figure real bequest where Social Security leaves zero. A worker in fixed income experienced the opposite — Social Security won by both measures — and the conservative-equity middle ground won on value while leaving four in five workers exposed to drawing less than Social Security in a bad market.

Stated limits

• No blended headline. How many workers would choose each band is a behavioral assumption this analysis deliberately does not make. All claims are per-band.
• P10-market income is a scaling approximation, pending a direct withdrawal-percentile read from the model.
• 5.6% of workers sit below the model's $20K income floor and are assigned floor-income outcomes.
• One cohort's actual history underlies the deterministic figures; the Monte Carlo figures, which resample it, are the horizon-robust measure — and every claim above survives on Monte Carlo alone.
• Survivor/disability benefits and taxation remain out of scope.

About these figures — please read. Every number here is an output of the Is Social Security Worth It? simulator, read from the August 2026 build. These are model estimates, not predictions, advice, or guarantees, describing the model on that date, not a promise about future versions. See Corrections & Retractions; figures may change in later builds. Past performance does not guarantee future results. Weights: Census FTYR median ($63,360, 2024); SSA taxable-max share (~8%); log-normal σ=0.73, stress-tested σ∈[0.60,0.90]; 2,000,000 draws, seed 42. Data: DCvsSSv3population.csv.