This page back-tests the payroll tax against the national average wage, all the way back to the first FICA cheque. Then it asks the obvious follow-up question — what if you had simply invested the money instead — and answers it honestly, including the parts of the answer that cut the other way.
How to read every figure here. The worker is a hypothetical who earns exactly the national average wage every year of a 45-year career. “FICA” means both halves — the employee's and the employer's — because economists generally treat the employer half as coming out of wages too. Dollars are converted to 2026 purchasing power. The nominal figures are always available in the table views.
The rate
2→15.3
The combined payroll tax rate went from 2.0 % of covered wages in 1937 to 15.3 % in 1990, and has sat there ever since. Nearly eight times the original rate.
OASDI plus Medicare Hospital Insurance. Medicare's share appears in 1966 and the rate has been frozen at 15.3 % since 1990.
Source: SSA, Social Security & Medicare Tax Rates (FICA & SECA). Excludes the 0.9 % additional Medicare tax on high earners introduced in 2013.
The bill, in today's money
The rate is only half the story. Real wages also grew, so the same percentage takes a bigger real bite. An average earner in 1951 handed over the equivalent of $1,048 a year in today's money. In 2026 it is $11,329 — and on the Trustees' own wage projections it reaches $13,664 by 2050.
Both halves, in constant 2026 dollars. Historical through 2025; the Trustees' projected wage path after that, at today's statutory rate.
Source: rates from SSA; average wage from SSA's AWI series and the 2026 Trustees Report table VI.G1; deflated with that report's CPI series spliced to CPI-U before 1970.
At the kitchen table
If your grandfather started work in the mid-1950s, he paid about a tenth of what you pay, in the same money, for the same programs.
He also got a far better deal on the way out, because he paid in for only part of a career at low rates and drew a full benefit.
In the committee room
The rate has not changed since 1990. Every increase in the real burden since then has come from wage growth, not legislation — which is why the tax has become more expensive without ever being voted on.
Closing both programs' gaps outright would take OASDI to 16.65 % and Medicare's HI to 3.46 % — a combined 20.11 %, a third above today.
The ratio you asked for
Same worker, same average wage, 45-year career — only the birth year changes. Everything is in 2026 dollars, so these are real comparisons, not the illusion of a bigger number.
All figures in constant 2026 dollars. Careers extending past 2025 use the Trustees' projected wage path at today's statutory rate — so if the rate ever rises to close the gaps, the later rows are understated.
The hypothetical
Take only the OASDI half — the retirement and disability part, not Medicare — and put every year's contribution into the S&P 500 at that year's actual total return. This is a back-test on real market history, not a projection.
All in 2026 dollars. Contributions are made at each year end and earn returns from the following year. Source: S&P 500 annual total returns (with dividends), 1928–2025, NYU Stern (Damodaran) from Federal Reserve and S&P data.
Identical contribution pattern; only the retirement date moves. Every end year the wage data supports, not a selection. This is sequence-of-returns risk, and it is why a single back-test proves very little.
Range across all 31 cohorts: 3.1× for a career ending in 2008, 9.0× for one ending in 2021 — a spread of nearly three to one. Same contributions, same market, different luck. Careers ending before 1995 cannot be computed because SSA's wage series begins in 1951.
The replacement cost
A career-average earner turning 62 in 2026 has a scheduled benefit of $2,608 a month — $31,301 a year, inflation-indexed for life. They pay $9,182 a year in OASDI tax to get it. So: what annual investment, over the same 45 years, would fund the same income?
Capital required at a 4% withdrawal rate: $782,537 in 2026 dollars. At 3% it is $1,043,382 and the S&P figure becomes $3,913 a year; at 5% it is $626,029 and $2,348. Returns are geometric and real, deflated by CPI. The denominator is the OASI share of the tax — 10.60 of the 12.40 combined points, or $7,849 of the $9,182 — because the benefit being replaced is a retirement benefit. Disability insurance is the other 1.80 points and buys something this comparison does not replace.
At the kitchen table
On the raw arithmetic, roughly 37 % of the retirement slice of your payroll tax, invested in a broad index over a full career at its long-run historical return, would have funded the same retirement income.
That gap is real, and it is the honest reason this argument never goes away. What the arithmetic leaves out is in the next section, and it is not small.
In the committee room
The comparison is a rate-of-return comparison, and a pay-as-you-go system cannot win one. Its return is roughly the growth of the wage base, which is far below equity returns by construction.
That is not evidence of waste. It is what happens when the first generation of a pay-as-you-go scheme receives benefits it never funded — every later cohort carries that cost forever.
The part that cuts the other way
Today's payroll taxes pay today's retirees. If workers redirected theirs into private accounts, the benefits already promised would still be owed — a $29.3 trillion present-value obligation that does not disappear because the funding did. You cannot pay the same dollar to a retiree and to a brokerage account.
Of the 70 million people Social Security paid in December 2025, about 8 million were disabled workers and their families and 6 million were survivors of workers who died. An index fund pays nothing extra if you are disabled at thirty or die at forty leaving children. That insurance is a large part of what the tax buys.
Social Security is a guaranteed, inflation-indexed income for life, with spousal and survivor protection. A 4 % withdrawal is a rule of thumb, not a guarantee — run out and there is no backstop. Buying an equivalent guarantee on the open market costs materially more than the figures above.
The same 45 years of contributions produced anywhere from 3.1× to 9.0× depending only on the year of retirement. Social Security's benefit does not depend on whether you turned 67 in 2009 or 2021.
That series is a price return — it excludes dividends, so it understates the index — but it also covers only 40 years containing a single extraordinary technology cycle, and the index is reconstituted to drop losers. Projecting 11 % real for 45 years forward is an extrapolation, not a finding. It is shown because it was asked for, not because it is a forecast.
And the obvious one. Nobody can actually opt out, so this is a thought experiment about program design, not a decision anyone gets to make. It is also not investment advice — I am not a licensed adviser, past index returns are not a promise of future ones, and the averages here conceal enormous variation between individuals. Social Security's formula deliberately replaces a much larger share of a low earner's wage than a high earner's; an average-wage comparison hides exactly that.