What one investment became
Each line is the same single investment, left alone. The scale doubles as it rises — that's what compounding looks like when you give it decades. Shaded bands are Canadian recessions; the labels mark moments that felt like the end of the world at the time.
The cost of panic
The hardest part of investing isn't picking investments — it's not selling them at the worst moment. This simulation sells whenever the portfolio falls 20% from its peak, waits in T-bills, and buys back a full year after the bottom. Pick which downturns would have shaken you.
The impact of time
One-year returns are a rollercoaster. Twenty-year returns, over the same lifetime, are almost boring. Switch between holding periods and watch the same history flatten — this is the single best argument for patience that data can make.
The full range, side by side
Each bar spans the worst to the best annualized return of every window of that length in your lifetime — Canadian stocks, US stocks and the balanced portfolio compared. Hover any bar to bring it forward and see its worst, average and best; click a portfolio in the legend to hide or show it. Periods longer than your window are left out.
What these terms mean
- Rolling return
- Instead of judging only calendar years, we look at every possible starting month in your lifetime. The 3-year rolling return for March 1998 is what you'd have earned buying in March 1995 and holding for three years. Plotting every window shows the complete range of what patience delivered — and what it didn't.
- Annualized
- All figures are expressed as a per-year rate so different holding periods can be compared fairly. A 10-year window that turned $10,000 into $20,000 is shown as about 7.2% per year, not 100%.
- Why longer windows flatten
- Great years and terrible years land inside the same long window and offset each other. Over one year the market's mood dominates; over ten, the underlying growth of businesses does. The range doesn't just narrow — historically, the worst long windows sit far closer to zero than the worst short ones.
Downturns and recoveries
Every fall of 20% or more in your lifetime — how deep it went, how long it fell, and how long it took to get back to even. The pattern to notice isn't the falls. It's the last column.
ⓘ How a downturn is measured
Depth is measured from the highest point the portfolio had ever reached — not from the start of a calendar year. A downturn only ends when the portfolio climbs all the way back above that old peak; rallies that fall short stay inside the same downturn.
That's why some episodes look surprisingly long. US stocks (in CAD) show August 2000 to February 2009 as a single downturn: in US dollars the market had recovered by 2007, but the Canadian dollar rose so much over those years that a Canadian investor holding US stocks was still below their 2000 peak when the financial crisis hit — one continuous episode, finally back to even in 2013.
Figures use month-end values, so the very worst single days fall between measurements — daily depths would be somewhat deeper. "Months falling" counts peak to bottom; "months to recover" counts bottom back to the old peak.
| Portfolio | Downturn | Depth | Months falling | Months to recover | Back to even by |
|---|
The chart is easy. The moment is hard.
Every line above only worked for people who stayed in their seat while the news was screaming. That's the real job of an advisor — not predicting markets, but keeping one bad month from undoing thirty good years.
Talk to De Thomas Wealth ManagementCommon questions
How often do market corrections and crashes actually happen?
More often than most people remember. In this dataset, Canadian stocks have had 21 corrections of 10% or more since 1956 — roughly one every three to four years — and 9 full bear markets of 20% or more. A diversified mix feels far fewer of them: the balanced portfolio crossed the bear-market line only 3 times in the same seventy years. The downturns table above lists every one that falls inside your own timeline.
How long does the market take to recover from a crash?
For Canadian stocks since 1956, the climb from the bottom back to the old peak has taken anywhere from 8 to 40 months, with a median around a year and a half. The outlier was a Canadian investor holding US stocks through the 2000s: because the Canadian dollar rose while the US market recovered, their 2000 peak wasn't regained until 2013. The recoveries column shows each episode in your window, and the ⓘ note there explains how the measurement works.
Is it better to sell during a downturn and buy back later?
History has been unkind to that strategy. The panic simulator above applies a common version of it — sell after a 20% fall, wait in T-bills, buy back a year after the bottom — and in this data it consistently ends with less money, because it misses the first year of each recovery, historically the strongest stretch. For a forward-looking version of the same question, our Monte Carlo simulator stress-tests a plan across thousands of scenarios.
What is a rolling return?
A rolling return measures every possible holding period of a given length, not just calendar years — the 10-year rolling return for March 2010 is what an investor earned buying in March 2000 and holding a decade. Plotting all of them shows the full range of what patience delivered. The impact of time section lets you switch between 1, 3, 10, 15 and 20-year windows across your own lifetime.
Does holding investments longer really reduce risk?
Historically, yes — dramatically. In this dataset, one-year outcomes ranged from deep losses to gains above 60%, but every single 20-year window since 1956, on Canadian stocks, US stocks and the balanced portfolio alike, ended positive; the worst 20-year stretch on the balanced mix still compounded at about 5% per year. Past results don't guarantee the future, but the pattern is the whole argument of the rolling-return charts above.
How does this connect to my actual retirement decisions?
This page shows what markets did; planning is about what you'll do next. For the decisions themselves, our other tools go deeper: when to start CPP, whether your savings will last with the retirement planner and Monte Carlo simulator, and what waiting to invest costs with the cost of waiting calculator. History is the evidence; those are the plans.
Methodology and sources
Where this data comes from
- Canadian stocks — S&P/TSX Composite total return, reconstructed from Statistics Canada's published index closes (v122620) and dividend yields (v122628). After StatCan discontinued the dividend-yield series in August 2016, the last published yield (2.9%) is held constant; after the index series ended in September 2023, month-end closes come from public market data (the two sources agree exactly over their 24-month overlap).
- US stocks — S&P 500 total return in Canadian dollars: Robert Shiller's public dataset (dividends throughout; prices to 1984), public month-end closes from 1985, converted at Bank of Canada / Statistics Canada monthly USD-CAD rates (v37426, then Bank of Canada Valet FXMUSDCAD).
- Bonds — a synthetic 10-year Government of Canada total-return index built from StatCan/Bank of Canada long-bond yields (v122487): buy a 10-year bond at par each month, sell it a month later, repeat.
- T-bills — 3-month Treasury bill auction average yields (v122484), compounded monthly.
- Balanced portfolio — 40% bonds, 30% Canadian stocks, 30% US stocks, rebalanced each January. A construction for illustration, not a recommendation.
- Inflation — Statistics Canada all-items Consumer Price Index (v41690973).
- Recessions — C.D. Howe Institute Business Cycle Council chronology (peak to trough months).
- Source links — Statistics Canada (vector numbers cited above), Bank of Canada Valet API, Robert Shiller's public dataset, and the C.D. Howe Business Cycle Council.
How the figures are calculated
- Growth chart — each line is a total-return index (dividends and interest reinvested) rebased so your starting amount is invested once, in January of your chosen year. The per-year figure on each card is the compound annual growth rate (CAGR) over your window.
- Rolling returns — every possible window of the chosen length is measured, sliding forward one month at a time, and each is expressed as an annualized rate. "Best" and "worst" are the extremes of those windows; "windows that made money" is the share that ended above zero. Windows overlap, so they are descriptive of history, not independent samples.
- Downturns — depth is measured peak-to-trough from the highest month-end value previously reached; an episode ends only when that old peak is regained (the ⓘ note in the downturns section walks through an example).
- Panic simulation — sells at the close of the first month 20% or more below the running peak, earns actual T-bill returns while out, and re-enters 12 months after that downturn's bottom. Overlapping episodes are merged. It is a fixed rule for illustration, not a model of any person.
- Balanced portfolio — 40% bonds, 30% Canadian stocks, 30% US stocks (in CAD), rebalanced each January.
- Inflation line — the CPI rebased to your starting amount: what that dollar figure would need to grow to just to keep its purchasing power.
Chart limitations
- These are index reconstructions from public data, not official index-provider figures. Validated against long-run published returns, annualized differences are small (a few tenths of a percent), but individual months can differ.
- Before 1985, US monthly figures are based on Shiller's monthly-average prices, which smooth month-to-month swings (long-run returns are unaffected). In particular, the best and worst 1-year rolling windows for US stocks understate the true extremes in any timeline that reaches back before 1985.
- No fees, taxes or trading costs are included, and you cannot invest directly in an index. Past performance does not predict future returns.
- The panic simulation is a rule, not a person: sell at 20% below peak, hold T-bills, rebuy 12 months after the bottom. Real behaviour is messier — usually in the expensive direction.