Open any retirement calculator and you’ll see the same default assumption: the stock market returns about 8 percent a year on average, sometimes 10. Plug in your contributions, watch the curve climb, retire happy. The number isn’t fabricated, but the way it’s deployed in planning software is misleading enough to generate a generation of underprepared retirees. The averages obscure more than they reveal.
Arithmetic average vs. geometric reality
The S&P 500’s long-run nominal arithmetic average return is in the high single digits, depending on the period measured. But arithmetic averages are not what your portfolio actually earns over time. The relevant figure is the geometric (compound) return, which accounts for the drag that volatility imposes. A portfolio that gains 50 percent one year and loses 33 percent the next has an arithmetic average of 8.5 percent and an actual compounded return of zero. Over the past century, the S&P’s geometric return has been roughly 6 to 7 percent nominally, and meaningfully lower after inflation. Retirement projections built on the higher arithmetic figure quietly inflate expected outcomes by 1 to 2 percentage points, which compounds into hundreds of thousands of dollars of phantom wealth over a 30-year horizon.
Sequence of returns matters more than averages
Even the corrected average misleads in retirement planning because the order in which returns arrive matters enormously once you start withdrawing. A retiree who hits a 30 percent drawdown in years one and two of retirement faces dramatically worse outcomes than one who experiences the same drawdown in year fifteen, even if the average return over both retirements is identical. This is sequence-of-returns risk, and it’s the single most underappreciated factor in retirement modeling. Monte Carlo simulations capture it; deterministic 8 percent calculators don’t. A retiree who withdraws 4 percent annually from a portfolio that delivers 8 percent on average but loses 30 percent in the first two years can run out of money a decade earlier than the average suggests.
What the number ignores entirely
Long-run return figures generally exclude or understate fees, taxes, and behavior. Even a 0.5 percent annual fee compounds into tens of thousands of dollars over decades. Taxes on dividends and capital gains in non-sheltered accounts further erode realized returns. And investor behavior — selling in drawdowns, chasing recent winners, mistiming rebalances — historically costs retail investors 1 to 3 percent annually compared to the index they hold, according to Dalbar and similar studies. Stack realistic geometric returns minus fees minus taxes minus behavioral drag, and the working assumption for planning purposes is closer to 4 to 5 percent real than 8 percent nominal. That’s a different retirement.
The bottom line
The 8 percent figure isn’t wrong in the narrow sense, but it’s used in ways that flatter projections and hide risks. Better planning starts with realistic compounded returns, accounts for sequence risk, and stress-tests against historical drawdowns. If a calculator or advisor builds a retirement plan on the 8 percent assumption without explaining its limitations, treat that as a signal about the quality of advice you’re getting, not a forecast.
Leave a Reply