Bookmark this now, while you're calm. A 30-year plan will live through four to six bear
markets, and each one will feel like the one that's finally different. This page exists
for those days: the full record of what crashes have done, and the short list of what
to do about them.
"Be fearful when others are greedy, and greedy when others are fearful."
The pattern repeats: a decline that felt world-ending at the time, followed by a recovery
that in hindsight looks inevitable. The table shows the S&P 500's major declines on a
price basis (dividends reinvested shortened every recovery below, in some cases by years):
Crash
Peak → trough
Decline
Back to prior peak
What it taught
Great Depression
1929–1932
≈ −86%
≈ 25 years (price)*
The worst case on record — and the origin of every rule about diversification and leverage. *With dividends and deflation-adjusted spending power, recovery took far less.
Stagflation bear
1973–1974
≈ −48%
≈ 7 years
Inflation is the enemy of both stocks and bonds at once; real returns are the ones that count.
Black Monday
1987
≈ −34% (−20% in one day)
≈ 2 years
A crash with no recession behind it can heal fast. Panic selling on the worst day was the only fatal mistake.
Dot-com bust
2000–2002
≈ −49% (Nasdaq ≈ −78%)
≈ 7 years (Nasdaq ≈ 15 years)
Concentration in one hot theme can lose a decade even inside a healthy economy. Broad beats narrow.
Global financial crisis
2007–2009
≈ −57%
≈ 5.5 years
The system itself can wobble — and steady monthly buyers through 2008–09 bought the best prices in a generation.
COVID crash
2020
≈ −34% in 33 days
≈ 5 months
The fastest bear ever, and the fastest recovery. Anyone who sold "until things calm down" missed it.
Inflation bear
2022
≈ −25%
≈ 2 years
Even a "mild" bear feels endless in real time — 10 months of falling prices tested every 2020-era investor.
Figures are approximate, S&P 500 price basis (Nasdaq noted where relevant), rounded for readability. Recovery = first close above the prior peak.
Read the table twice. Once for the declines — that column is why your
allocation must match your stomach. Then again for the recovery column: every single
crash in a century ended the same way, with the market at new highs. The people hurt
permanently were almost never the ones who held; they were the ones who sold at the
bottom, or held leverage and never got the chance to decide.
The playbook
What to actually do in a crash
Do
Nothing, mostly. If your plan was right last month, a lower price doesn't make it wrong. Doing nothing is an active, correct decision.
Let the automation run. Your monthly buy is now purchasing more shares per unit of currency than it did at the peak. This is the mechanism working, not failing.
Re-read your investment policy — the one-page document you wrote when calm (see Principles). It was written precisely for today.
Rebalance if your bands say so. A crash pushes stock weight below target; rebalancing then means buying low on a rule, not on a feeling.
Check your cash runway, not your portfolio. The question that matters is "can I avoid selling for the next year or two?" — not "how much is it down today?"
Don't
Don't sell "until things settle down". Recovery days cluster inside the chaos; missing a handful of them costs a large share of lifetime returns.
Don't check prices daily. Nothing in a 2050 plan requires today's quote. Every check is another chance to panic.
Don't reach for leverage or "buy the dip" money you may need. Crashes get deeper than feels possible; only invest cash with a decade-long runway.
Don't read forecasts. Nobody called the bottom in 1932, 1974, 2009, or 2020. The confident voices on the way down were guessing too.
Don't confuse your market with every market. A crash in one country is a reminder to stay diversified, not a reason to abandon the plan.
Why holding works
The arithmetic of panic
Three facts, all covered in depth on the Principles page,
do the heavy lifting:
The best days live inside the worst months. Studies of the S&P 500
repeatedly find that missing only the 10 best days across decades roughly halves final
wealth — and those days overwhelmingly occurred during bear markets, next to the worst days.
A falling market raises the return on every new contribution. For anyone
still adding money monthly, a bear market is the sale of the decade running in slow motion.
The investors who compounded through 2008–09 or March 2020 bought their best-performing shares then.
Declines are temporary for broad markets, permanent only for the leveraged and
the concentrated. Individual companies go to zero; the S&P 500 never has. That
asymmetry is the entire case for owning the broad market instead of a handful of names.
The honest exception. "It always comes back" has one documented failure mode:
a single market bought at bubble valuations — Japan needed 34 years (see
the Japan case study). The insurance is the same as always:
broad diversification, steady buying rather than one lump at a peak, and no leverage.
A global or US-broad portfolio bought monthly has no historical precedent for permanent loss
over 20+ years. That is a statement about history, not a guarantee — but it is the strongest
base rate in investing.
Before the next one
Crash-proof the plan while the sun is out
Hold 3–6 months of expenses in cash. A job loss during a bear market is
the one scenario that forces selling at the bottom. The emergency fund exists so the
portfolio never has to be one.
Choose an allocation whose worst year you can watch. Rough rule: your
stock percentage, halved, is a loss you should expect at least once a decade. 80% stocks
→ be ready to see −40% on a statement without selling.
Write the policy now. One page: what you own, why, what you will do at
−20% and −40% (typically: nothing, or rebalance into stocks). Signed and dated. Future-you
will not believe past-you was so calm.
Automate everything so that continuing correctly requires no courage —
and stopping requires effort.
Pre-decide your information diet: in a crash, check the portfolio
monthly at most, and let the annual review (see Strategy)
stay annual.