S&P 500 · Rolling Return Instrument
The same index, the same monthly closes, run through ten different rolling windows — from six months to twenty-five years. Each panel below re-annualizes the trailing return at every point in time, so a short window jitters with every scare and rally while a long window settles into the market's slow, underlying drift. Click any panel to bring it into focus, or layer several on top of each other to see where the horizons agree — and where they don't.
Each series is the S&P 500's total price change over the trailing window, converted to a
compounded annual rate: ((Pₓ / Pₓ−w)^(1/years) − 1) × 100.
That puts every horizon on the same footing — a 6‑month window and a 25‑year window
both read out in "percent per year," even though one is dominated by the latest headline and the
other has lived through several full market cycles. Values below zero mean the index was lower,
annualized, than where that window started. Flip Inflation‑adjusted above to
run the same calculation on CPI‑deflated (constant‑dollar) levels instead of nominal
ones — that's the difference between what the index said you made and what it was
actually worth in today's purchasing power.
Click a panel to load it into the instrument above.
What these numbers do and don't tell you.
An ordinary return measures one fixed start date to one fixed end date. A rolling return recalculates that same measurement at every point in time — so instead of "the market returned 8% from 2010 to 2020," you see what a 10-year holding period returned if you'd started in January 2010, then February 2010, then March, and so on. It answers a more useful question: not "what happened," but "what would have happened to me depending on when I happened to start?"
Partly because they genuinely are — good and bad years offset each other — and partly because annualizing compresses them. A 60% crash in one year is a 60% drop on the 1-year line, but spread across a 25-year window it becomes a couple of percentage points per year. Both effects are real, but it's worth remembering the calm of the 25-year line is partly a measurement artifact, not just market serenity.
No. These are price returns only. Historically dividends have added roughly two percentage points per year, so a total-return version of this page would sit meaningfully higher — particularly on the long horizons, where that gap compounds. If you're comparing these numbers to a fund's reported performance, that fund almost certainly includes reinvested dividends and will look better as a result.
It swaps the underlying price series for a CPI-deflated one, so returns are expressed in constant purchasing power rather than nominal dollars. The gap is not small: over the trailing decade it's worth several percentage points a year. Nominal returns tell you what your statement said; real returns tell you what you could actually buy.
Robert Shiller's long-run dataset, which carries monthly S&P 500 prices and the matching Consumer Price Index back to 1871 in a single file — so the nominal and real series stay internally consistent. This page checks for new data daily and rebuilds itself when a new month is published. The footer shows the latest month included.
No. It's a historical data visualization. Past returns describe what happened to a specific index over a specific stretch of history; they are not a forecast, and nothing here accounts for your taxes, fees, timeframe, or circumstances.