Bear Markets — 100 Years of History (Marcin Iwuć)
🗒️ Description
A synthesis of Marcin Iwuć’s notes (marciniwuc.com, Finansowa Forteca, CFA) on bear markets over the last ~100 years. Core scope: the US market (S&P 500), with a global accent (MSCI ACWI / MSCI World). Data as of 14 June 2026.
Central thesis: “You cannot predict, but you can prepare” (Howard Marks). Times and technologies change, but human psychology does not — and it is people’s emotional decisions that ultimately decide whether prices fall or rise. Every bear market looks different, yet they all rhyme. They cannot be eliminated — only prepared for.
The original is the author’s personal “scratchpad” used for his own investment decisions. Educational only — not investment advice. When the author writes “do X,” it is a message to himself.
🔗 Links
- Part I — Definitions and Frequency
- Part II — Historical Tables
- Part III — The Five Phases of a Bear Market
- Part IV — Indicators — a Compass, Not a Clock
- Part V — Where We Are Today (June 2026)
- Part VI — Thinking About Risk — Scenarios and Stress Test
- Part VII — Common Analytical Errors and Narrative Traps
- Glossary
🧩 Part I — Definitions and Frequency
Three thresholds of decline
| Term | Threshold (from peak) | Character | Frequency (US) |
|---|---|---|---|
| Correction | drop ≥ 10% | frequent, usually short | ~every 1.2 years (since 1980) |
| Bear market | drop > 20% | rarer, deeper | ~13–15 post-war; every ~4–5 years |
| Crash | sharp drop ≥ 30% in short time | rare, panicked | a handful (1929, 1937, 1987, 2008, 2020) |
Source: Hartford Funds / Ned Davis Research, “10 Things You Should Know About Bear Markets”, 2025. Goldman Sachs (“Bear Market Anatomy”, 2025) counts 27 bear markets on the S&P 500 since 1835, 10 post-war.
The 20% threshold is arbitrary. The 1990 bear stopped at −19.9% on closing prices (intraday it crossed 20%) and is classified both ways. Always check whether a figure refers to closing vs intraday, nominal vs real, price index vs total return — these differences explain most discrepancies between sources.
Why bear markets repeat: the credit cycle
Bear markets are not random misfortunes — they are baked into the mechanics of a credit-based economy. Two overlapping cycles matter:
- Short-term credit cycle (business cycle, ~5–10 years). Economy grows → banks lend freely → credit fuels spending → asset prices rise. Inflation appears → central bank hikes rates → credit shrinks → spending falls → slowdown/recession with a cyclical bear. Imbalances clear, cycle restarts. This is why bears arrive every few years.
- Long-term debt cycle (~50–75 years). Over decades debt grows faster than income; each short cycle leaves higher debt. When debt service becomes too heavy, large-scale deleveraging happens — and that is when structural bears occur (1929, 2008).
Ray Dalio calls this the “big debt cycle”; Hyman Minsky’s thesis: “stability breeds instability” — a long calm encourages ever more risk-taking until the system is fragile (the Minsky moment).
Key takeaway: the reasons for bears change, but the mechanics of credit and human psychology do not.
🧩 Part II — Historical Tables
Table 1. Post-war S&P 500 bear markets (closing prices, peak→trough)
| # | Peak | Trough | Main trigger | Decline | Time to fall | Time to recover | Recession? |
|---|---|---|---|---|---|---|---|
| 1 | 1946-05-29 | 1949-06-13 | post-war adjustment | −29.6% | ~36.5 mo | ~4.1 yr | yes |
| 2 | 1956-08-02 | 1957-10-22 | tightening, ‘57–58 recession | −21.6% | ~14.7 mo | 11 mo | yes |
| 3 | 1961-12-12 | 1962-06-26 | ”Kennedy Slide” | −28.0% | ~6.5 mo | 14 mo | no |
| 4 | 1966-02-09 | 1966-10-07 | credit crunch, Vietnam | −22.2% | ~8.0 mo | 7 mo | no |
| 5 | 1968-11-29 | 1970-05-26 | inflation, tightening | −36.1% | ~17.8 mo | 21 mo | yes |
| 6 | 1973-01-11 | 1974-10-03 | oil embargo + stagflation | −48.2% | ~20.7 mo | ~5.8 yr | yes |
| 7 | 1980-11-28 | 1982-08-12 | Volcker vs inflation | −27.1% | ~20.4 mo | 3 mo after trough | yes |
| 8 | 1987-08-25 | 1987-12-04 | crash (portfolio insurance) | −33.5% | ~3.3 mo | 1.7 yr | no |
| 9 | 1990-07-16 | 1990-10-11 | Gulf War, recession | −19.9% | ~2.9 mo | 4 mo | yes |
| 10 | 2000-03-24 | 2002-10-09 | dot-com bubble burst | −49.1% | ~30.5 mo | ~4.7 yr | yes |
| 11 | 2007-10-09 | 2009-03-09 | global financial crisis | −56.8% | ~17.0 mo | ~4.1 yr | yes |
| 12 | 2020-02-19 | 2020-03-23 | COVID shock | −33.9% | 33 days (fastest) | 5 mo | yes |
| 13 | 2022-01-03 | 2022-10-12 | inflation + Fed tightening + Ukraine/energy shock | −25.4% | ~9.3 mo | ~24.5 mo | no (no official recession) |
Average (13 bears): −33.2% · Median: −29.6%
Sources: Yardeni Research; simianx “Every S&P 500 Bear Market Since 1929”; Wikipedia (cross-check).
Table 2. Extremes / “long tail” (separate case studies — NOT in the averages)
| Period | Trigger | Drawdown | Recovery | Note |
|---|---|---|---|---|
| 1929–32 (Dow) | 1920s bubble, Fed errors, crash | −89% (381→41) | 25 yr nominal (to Nov 1954) | canonical tail; start of the Great Depression |
| 1907 (Panic) | bank run, Knickerbocker Trust | Dow ~−50% in weeks | — | case study |
| 1937–38 | premature tightening | ~−54%+ | ~9 yr | bear within the Depression decade |
On 1929: the famous “25 years underwater” refers to the Dow (price index, nominal, no dividends), recovered Nov 1954. For the S&P composite the nominal price return came back faster but still over a decade. Deflation 1930–32 actually helped (real losses < nominal), and dividends were high (4–6%/yr), shortening recovery. Per Shiller/DQYDJ, an investor who bought at the very Sept-1929 peak and reinvested dividends recovered in real terms in ~7–7.5 years — around 1936/37, right before the 1937–38 recession pushed them back down.
Table 3. Real S&P Composite drawdowns (total return, since 1871) — bears ≥ −20%
Measure: real total return (after inflation, with dividends) — the truest picture of the investor’s experience. Median drawdown −31.7%; median peak→trough 18 mo; median recovery from trough 24 mo.
| Peak | Trough | Drawdown | Recovery from peak | CAPE at peak |
|---|---|---|---|---|
| Sep 1929 | Jun 1932 | −76.8% | ~86 mo | 32.6 |
| Aug 2000 | Mar 2009 | −51.8% | ~153 mo | 42.9 |
| Jan 1973 | Dec 1974 | −50.1% | ~144 mo | 18.7 |
| Feb 1937 | Apr 1942 | −48.3% | ~98 mo | 22.2 |
| Nov 1916 | Dec 1920 | −47.1% | ~93 mo | 12.1 |
| Sep 1906 | Nov 1907 | −36.7% | ~32 mo | 19.2 |
| Apr 1946 | Feb 1948 | −35.4% | ~54 mo | 16.0 |
| Dec 1968 | Jun 1970 | −31.7% | ~47 mo | 22.3 |
| Aug 1987 | Dec 1987 | −26.7% | ~24 mo | 18.3 |
| Nov 2021 | Oct 2022 | −24.5% | ~28 mo | 38.6 |
| Dec 1961 | Jun 1962 | −21.8% | ~17 mo | 22.0 |
Source: Robert J. Shiller, Online Data (ie_data.xls), own calculations.
In real-total-return terms, the dot-com (2000) and 2008 crises merge into ONE ~13-year submersion (2000 peak only durably exceeded in 2013) — the “lost decade” lasted longer than 10 years.
Why the framing matters (price vs total return, nominal vs real). Most bear-market headlines use the price index, nominal — the most dramatic version. But (1) reinvested dividends account for a large share — in the long run roughly half — of total equity return; the price index ignores them, overstating recovery time. (2) Inflation/deflation: the real view can flip the picture (1929 = ~7 yr real vs 25 yr nominal; the 1970s = longer real recovery despite nominal bounces). When someone scares you with a bear-market number, ask three questions: which index, price or total return, nominal or real?
Table 4. Three species of bear (Goldman Sachs)
| Type | Mechanism | Avg decline | Avg duration | Avg recovery |
|---|---|---|---|---|
| Structural | imbalances + bubble burst / banking crisis | ~−57% (>50%) | ~42 mo (~4 yr) | ~8–10 yr |
| Cyclical | business cycle: rates up, recession, falling earnings | ~−31% | ~21–27 mo | ~50 mo (~4 yr) |
| Event-driven | one-off shock (war, oil, pandemic) | ~−29% | ~8–9 mo | ~12 mo (~1 yr) |
Source: Goldman Sachs, “Global Strategy Paper: Bear Market Anatomy” (Peter Oppenheimer & team), 2025-04-08.
Recessionary vs non-recessionary (CFA Institute, 15 bears since 1950): median decline −35% (with recession) vs −22% (without); median duration ~18 mo vs ~3 mo. Of 11 rate-hike cycles: 9 yield-curve inversions, 8 recessions with a bear; the only exception (inversion without recession) was 1966.
Counter-trend rallies in bear markets (bear-market rallies)
Goldman Sachs, “Bear Market Anatomy”, Exhibit 7 — 19 global episodes (MSCI ACWI). Average 44 days / +14.1% (median 46 days / +12.8%). Cyclicals beat defensives in 83% of rallies; emerging beat developed in 67%. Each is a rally “from local trough to local peak” inside an ongoing bear — after which (except the last, true bottom) the market fell lower.
Analytical takeaway. A bear doesn’t fall evenly — it falls in zigzags. The GFC 2008 had six counter-trend rallies (two of ~+22% and +24%). The ten best single days for the S&P 500 in history fell in 1987, 2008, and 2020 — many of them inside these rallies. This is the core argument against timing: leaving the market to dodge declines very likely also misses the best days, which drive most of the long-term result.
Iwuć’s conclusion: you can’t tell a true bottom from a false one live — the difference is only visible in hindsight. (Marks: “there’s no way to know for sure whether a rise was justified or irrational.”)
Global dimension (MSCI World) — the “global lost decade”. The worst MSCI World drawdown is one continuous episode Sep 2000 → Mar 2009 (~−57.8% USD; −55.7% for a euro investor, full recovery only May 2014 — nearly 14 years), because the global index never reclaimed its 2000 peak before 2008 hit. In the US narrative these are two separate bears; globally it is one multi-year lost decade. But global shocks hit everywhere alike — in COVID 2020 MSCI World fell ~−34%, almost identical to the S&P 500. Geographic diversification protects against local problems, not against a simultaneous worldwide collapse.
Banking crises — the heaviest category (Reinhart & Rogoff, “The Aftermath of Financial Crises”, NBER WP 14656, 2009). When a bear is accompanied by a financial-system rupture:
- real equity decline: avg −55.9% over a 3.4-year downturn;
- real housing decline: −35% over ~6 years;
- unemployment rises ~7 percentage points;
- GDP falls ~−9.3%;
- public debt rises ~86% (mostly from collapsing tax revenue, not bank bailouts).
This is the difference between a cyclical bear (−31%, ~2 yr) and a structural one (−57%, ~decade): the latter is defined by banking-system involvement. Hence the weight given to credit spreads in the “compass”.
🧩 Part III — The Five Phases of a Bear Market
The reasons change (banks, oil, virus, AI), but the sequence of emotions and behaviors is repeatable, because it stems from human psychology.
Why psychology beats knowledge. Cognitive traps that fire across the cycle: loss aversion (pain of loss ~2× the joy of gain — fear phase); herding / informational cascades (euphoria + capitulation); overconfidence (mistaking a bull market for talent — euphoria); recency (extrapolating recent experience — all phases); anchoring (to the last peak or trough — denial, recovery); disposition effect (holding losers, selling winners); mental accounting; snakebite effect (avoiding a whole asset class after pain — capitulation, recovery). You can’t switch these off — only design around them with portfolio construction and rules. That is the meaning of “tying yourself to the mast”: not heroic willpower, but a system that doesn’t require you to make the right decision while panicking.
Phase 1. Euphoria (cycle peak)
- Market/macro: index at all-time highs; stretched valuations (high CAPE, high Buffett Indicator); narrow leadership; rising imbalances (record margin debt, IPO/SPAC wave, speculative assets rising fastest). Economy looks strongest, but leading indicators start to weaken and policy tightens.
- Psychology: Marks’s pendulum at maximum greed — the paradox: maximum financial risk, while everyone subjectively feels safest. Herding, overconfidence (most active traders earned 11.4%/yr while the market gave 17.9% — Barber & Odean), recency, “this time is different”.
- Anatomy of mania (Minsky/Kindleberger): (1) displacement → (2) boom → (3) euphoria → (4) distress (smart money quietly exits) → (5) panic/revulsion. Euphoria is the second-to-last stage before the break, not the peak itself.
- Examples: 2000 dot-com (CAPE record 44.2; “Dow 36,000” bestseller); 2008 (margin-debt peak ~3 mo before market peak); 2020 (no classic euphoria — event-driven).
- Do: hold target weights; rebalance (automatically trim what grew beyond plan); write the bear-market plan before it starts; build cash cushion/ammunition; reduce leverage and liquidity risk.
- Don’t: add leverage at the top; chase the asset that “keeps rising”; mistake the bull for talent; buy only because it recently paid off.
- Advice: euphoria isn’t about “escaping at the top” (you can’t recognize the top live) — it’s about letting rebalancing mechanically take some chips off the most expensive asset, without emotion or forecast.
Phase 2. Denial (start of the bear)
- Market/macro: first leg down, index slides 10–20%; volatility rises but the narrative calls it a “healthy correction / buy the dip”. Most corrections do stop here — which reinforces denial.
- Psychology: anchoring (the recent peak as “the value the market will return to”), recency, cognitive dissonance.
- Examples: 2000 (rallies +9% to +20% fed “the bottom is in”); 2008 (Jan/Mar 2008 rallies +8%/+14%); 2020 (denial barely happened — euphoria to panic in weeks).
- Do: stick to the plan; execute scheduled buys/rebalancing mechanically in tranches; verify real risk (leverage, liquidity, concentration).
- Don’t: average down chaotically; add risk assuming “just a correction”; freeze waiting only for a return to the peak.
- Saying: “it’s just a correction” is true 3 times out of 4 — and costly in the fourth. Don’t guess which; hold the plan.
Phase 3. Fear (main decline)
- Market/macro: decline deepens; with a recession/financial crisis the drawdown exceeds 30%+ (recessionary median −35%, ~18 mo). Earnings revised down, spreads widen, liquidity worsens. The sharpest drops AND rebounds happen near the bottom — this is where the counter-trend rallies appear.
- Psychology: loss aversion rules (Kahneman & Tversky 1979) → cutting positions “just to survive”, usually at the worst moment. Herding down, market overreaction (De Bondt & Thaler), snakebite effect.
- Examples: 2008 (post-Lehman 15 Sep 2008 panic; VIX 80.86 on 20 Nov 2008; Buffett “Buy American. I Am.” 17 Oct 2008, Marks’s “Nobody Knows” memo — both bought ~5 months before the March-2009 bottom); 2020 (VIX 82.69 on 16 Mar 2020, all-time record).
- Do: continue the buying plan (DCA) and rebalancing (= buying discounted equities); monitor the whole portfolio as one organism; return to the written plan and execute it; treat the decline as a sale of assets you intended to accumulate long-term.
- Don’t: cut positions out of pain; flee entirely to cash hoping to “wait and re-enter at the bottom” (you’ll miss the best rebound days); treat counter-trend rallies as the all-clear; fixate only on the losing equity sleeve.
- Advice: in the fear phase your only real edge is the earlier decision. If the plan was written in euphoria, here you just execute it. If there’s no plan — fear is the worst time to create one.
Phase 4. Capitulation (the bottom)
- Market/macro: maximum pessimism. Panic, forced selling (margin calls, fund redemptions), media narrative “the system is collapsing / stocks are over”. The market is cheapest exactly when nobody wants to buy. Paradox: maximum opportunity precisely when it’s emotionally hardest to buy. The best days cluster right next to the bottom. The bottom is invisible live.
- Mechanics of forced selling (liquidity doom loop): (1) margin calls force leveraged sellers; (2) fund redemptions force managers to sell; (3) deleveraging — institutions cut risk. Each sale lowers prices → more calls/redemptions. That’s why bottoms are so violent and “overshot”: at the end, those who must sell dominate those who want to. Whoever does not have to sell (cushion, no leverage) is privileged — they can buy from the forced.
- Psychology: extreme recency (extrapolating catastrophe forever), full-force snakebite (some quit the market for years), pendulum maxed toward fear.
- Examples: 2008 (AAII 70.3% bears on 5 Mar 2009, days before the bottom; S&P 500 trough 676.53 on 9 Mar 2009; market then +56.9% in 12 mo); 2020 (bottom 23 Mar; “+17.6% in 3 days” — biggest 3-day gain in 80+ years).
- Do: don’t sell at the bottom; if you have a plan and cash, execute the buys/rebalancing (best long-term buys are made here); stay tied to the mast; treat sentiment extremes as context (“close”), not a date signal.
- Don’t: sell “everything” and swear off stocks; try to pick the exact bottom (even Buffett and Marks couldn’t); extrapolate catastrophe forever.
- Saying: maximum opportunity arrives disguised as maximum fear — most don’t seize it not because they can’t see it, but because they can’t act against their own emotions.
Phase 5. Rebound and recovery
- Market/macro: market rises — often sharply and early — against still-bad headlines (“climbing a wall of worry”). The first leg of the new bull is fastest and most distrusted. Macro still weak (unemployment often rises after the market bottoms — the market leads the economy by months). Recovery to the prior peak: months (shallow/event-driven) to years (deep/structural).
- Psychology: disbelief (“just another false dawn”), pessimistic recency, anchoring shifts to the bottom. Retail returns only when prices are already much higher.
- Examples: 2000 (~4.7 yr to peak, by 2007); 2008 (~4.1 yr, by 2013); 2020 (fastest “V” ever, ~5 months).
- Do: be in the market (the first, strongest rebound is only caught if you never left); continue plan/rebalancing; update the plan for the next cycle (Phase 5 eventually becomes Phase 1).
- Don’t: wait “until it’s safe” (safety comes with much higher prices); try to “jump in at the perfect moment” after exiting — that’s double timing, which fails both ways.
- Advice: the biggest cost of a bear isn’t the decline — it’s the missed rebound. Whoever sold in fear or capitulation usually returns only in Phase 5, paying the highest possible price for peace of mind.
The Four Storms in full: (1) Dot-com 2000–02 — longest capitulation; the most narrative manias leave the deepest scars. (2) GFC 2007–09 — textbook structural/banking bear; textbook gap between right direction and elusive timing. (3) COVID 2020 — fear+capitulation compressed into weeks; proof the calendar is unpredictable but the sequence is repeatable. (4) Great Depression 1929–32 — the −89% long tail; an extreme, not a template — yet proof the market eventually emerges, and that “25 years” is a nominal myth (~7 yr real with dividends).
🧩 Part IV — Indicators — a Compass, Not a Clock
Indicators don’t predict the moment — they gauge which phase of the cycle we’re roughly in. Overriding rule: extremes are contrarian statistically and with a lag; no indicator gives a date. Timing legend: W = leading, K = coincident, O = lagging, ⊘ = non-timing (valuation/context).
| Indicator | Euphoria | Denial | Fear | Capitulation | Rebound | Timing |
|---|---|---|---|---|---|---|
| Yield curve 10Y–3M | inverted / steepening | steepened | positive | positive | positive | W |
| HY OAS spreads | tight (<350bp) | widening | wide | extreme (800–2000+bp) | tightening | W |
| ISM / PMI | >50, weakening | ~50 / <50 | <50 | trough (~32 in 2008) | rebounds >50 | K |
| % stocks > 200 DMA | ↑ >80% (or weakening) | falling | low | <15% | thrust ↑ | K |
| Sahm Rule | <0.5 | ~0 | triggers | >0.5 | falling | O |
| Buffett Indicator | extreme | high | falling | lower (deep bears) | rebounds | ⊘ |
| Shiller CAPE | extreme | high | falling | lower (deep bears) | rebounds | ⊘ |
| AAII Bull-Bear | bulls ↑ | bulls falling | bears rising | bears extreme (~70%) | bulls rise late | contr. |
| VIX | low (<15) | elevated | 30–50 | ~80 (panic) | falling | K |
| CNN Fear & Greed | Extreme Greed | Fear | Fear/Extreme Fear | Extreme Fear (0–10) | back toward Greed | contr. |
| GS Bull/Bear Indicator | high (>70%) | high | falling | low (<40%) | low, rising | ⊘ |
| Equity fund flows | inflows peak | inflows weaken | outflows rise | redemptions extreme | inflows return (late) | contr. |
Readings at turning points (peaks vs bottoms):
| Indicator | Typical at peak | Typical at bottom (example) |
|---|---|---|
| Shiller CAPE | 44.2 (Dec 1999); 41.4 (2026) | ~13 (Mar 2009); real ~5–7 (1932, 1982) |
| Buffett Indicator | ~146% (2000); ~231% (2026) | ~56% (2009) |
| AAII bulls / bears | bulls ~75% (2000) | bears 70.3% (5 Mar 2009) |
| VIX | <13–15 (euphoria) | 80.86 (2008); 82.69 (2020) |
| HY OAS | 2.41% (Jun 2007, record min) | 21.82% (Dec 2008, record max) |
| ISM (mfg) | >55 (expansion) | 32.4 (Dec 2008) |
Confluence beats any single gauge: the strongest signal comes when several independent axes (valuation, credit, sentiment, breadth) light up at once. A single extreme reading is a curiosity; four together describe a phase. But even confluence gives no date.
Misleading indicators — flashy but unreliable:
- Margin debt — sounds like the ideal “too much leverage” signal, but it’s coincident (correlation with next-month return ≈ 0.00). Keep it as a description of euphoria, not a signal.
- “Best days” / “behavior gap” one-sided — “miss the 10 best days, lose half your gains” is abused as “never sell”. Honestly: best and worst days cluster in the same turbulent periods — they travel in pairs. The behavior gap “~1.2%/yr” (Morningstar/DALBAR) was challenged in peer-reviewed work (Financial Analysts Journal, 2026) — don’t use it.
Saying: the barometer says “low pressure”, not “storm at 2 PM”.
🧩 Part V — Where We Are Today (June 2026)
Snapshot, deliberately dated — a photo of a moment, not a forecast. Verify FRED-based figures directly at fred.stlouisfed.org before use.
| Indicator | Value | Date | Historical context |
|---|---|---|---|
| Shiller CAPE | 41.4 | 2026-06-12 | avg 17.4; record 44.2 (Dec 1999); ~96th percentile |
| Buffett Indicator | ~231% | 2026-06 | record; ~40% above long-term avg |
| Yield curve 10Y–3M | +~0.8 pp (positive) | 2026-06-11 | after 2022–24 inversion; un-inversion often precedes recession |
| HY OAS spreads | ~2.78% | 2026-06 | very tight; record min 2.41% (Jun 2007) |
| Berkshire Hathaway cash | $397.4B | Q1 2026 (pub. 2026-05-02) | all-time record; up from $373B |
Interpretation (no timing). A late-cycle picture of elevated risk: expensive stocks (CAPE near the dot-com record), extreme Buffett Indicator, tight spreads (complacency), freshly steepened curve, record smart-money cash. Yet the curve and spreads do not yet signal recession — a typical “expensive but unbroken” market.
- Valuation & sentiment group (red): CAPE 41.4, Buffett ~231%, record Berkshire cash — non-timing/contrarian. Message: future 10-year returns are likely low with a small margin of error. What they DON’T say: when (can persist for years).
- Cycle & credit group (green/neutral): positive curve, very tight HY (~2.78%), no recession signal — leading indicators. Message: no recession trigger visible yet — and recession decides whether a bear is shallow (−22%) or deep (−35%+).
Conclusion: 2026 is “expensive, but not breaking” — exactly how euphoria/late-cycle looked in 1999, 2007, and 2021, sometimes long before the peak. Elevated risk justifies discipline (weights, cushion, no leverage), not exiting the market awaiting a date no one knows. A recession trigger (widening HY, ISM collapse, jump in unemployment) would turn the cycle/credit group red — a moment for heightened vigilance, still not a date forecast.
🧩 Part VI — Thinking About Risk — Scenarios and Stress Test
Since the moment can’t be predicted, think in scenarios: not “when” but “what am I prepared for”. Three variants of a future bear:
- Event-driven (external shock — geopolitics, oil, pandemic): ~−25/−30%, short, fast recovery — unless it turns into recession. Pattern: 2020.
- Cyclical (recession — the default, most likely variant): ~−30/−35%, ~18 mo, recovery ~2–4 yr. Pattern: 1990, 2022 (no recession — shallower).
- Structural (systemic crisis with leverage/credit): ~−50/−57%, recovery in years (globally ~a decade). Pattern: 2008, 2000. Reinhart–Rogoff territory (−55.9% / 3.4 yr).
Don’t guess which scenario comes. Build a portfolio that survives all three — differing only in how much it hurts. That is the point of diversification and a cushion: not to maximize gain in one scenario, but to stay in the game in all of them.
The trap of betting on the worst. Building the whole strategy around the structural scenario (−57%) is costly — most bears are milder (−22% to −35%); sitting in cash “waiting for −57%” usually means missing multi-year bulls (the Grantham 2022 case). Keep the extreme in mind as a tail, not a base case.
Portfolio stress test (do this exercise). Take your current equity value and compute what’s left after −22% (non-recessionary median), −35% (recessionary median), −50% (structural). Then answer honestly: (1) Can I hold emotionally without selling? (2) Will I be forced to sell (no cushion, loan payments, income loss)? (3) Do I have ammunition (cash) to buy more? Any “no” means your equity weight may be too high today — not that “you must wait for a crash”.
Bear-market plan — checklist. A written plan is the mast you tie yourself to. Include: target asset-class weights (and max acceptable total-portfolio drawdown); cushion size (months of expenses outside equities); rebalancing rule (deviation threshold + frequency, e.g. quarterly or ±5pp); contribution rule (DCA — fixed amount at fixed intervals, regardless of headlines); leverage rule (ideally zero in the long-term portfolio); pre-mortem — write explicitly what you will and won’t do in fear and capitulation, so you only execute in panic, never create.
🧩 Part VII — Common Analytical Errors and Narrative Traps
- “This time is different” — the most expensive four words in finance (Reinhart & Rogoff). Spoken in euphoria (“new economy justifies any valuation”) and in capitulation (“the market won’t recover this time”). Always an extrapolation error.
- “Bear porn” — fear gets clicks. Crash-predicting content earns attention, so media/creators have a structural incentive to produce it regardless of market state. Separate signal (hard data) from noise (headlines).
- One axis posing as a chorus. One sensible thesis (e.g. expensive AI stocks) quoted by a hundred commentators creates “everyone knows” — not a hundred independent proofs, but one repeated a hundred times. Ask: how many independent risk axes do I see?
- Confusing a return predictor with a timing signal. CAPE/Buffett forecast 10-year returns, not the moment. “CAPE at a record, so sell” is a category error.
- Survivorship & selective guru citation. We hear of those who called it (Burry 2008), not the hundreds who warned and were wrong. Even the right-once usually miss repeatedly (Grantham). One correct call ≠ predictive ability.
- Recency & anchoring. After a long bull, declines feel unreal; after a deep bear, rallies feel like traps.
- Overfitting on a small sample. US crashes (>30%) are few — easy to “find” patterns that aren’t there.
- Narrative fallacy (Taleb) / hindsight bias. After the fact every crash has an “obvious” cause and a trail of warnings that “were visible”. In real time those same signals drowned in the noise of many others that led nowhere.
Iwuć’s note: the hardest analytical discipline isn’t finding bearish arguments — there are always plenty. It’s honestly weighing them against the case for the bull continuing, aware that your own psychology (fear, need for control, desire to “be right”) tips the scale. Balance of evidence, not advocacy for a thesis.
🧩 Glossary
- Drawdown — decline from last peak to trough, in %. Max drawdown = the largest such decline in a period.
- Correction / bear / crash — drop ≥10% / >20% / sharp ≥30% respectively.
- Price index vs total return — price index counts only price change; total return adds reinvested dividends. Long-term the difference is huge.
- Nominal vs real — real = inflation-adjusted (purchasing power). The same data can give a different recovery-time picture (see 1929).
- CAPE (P/E10, Shiller PE) — price to 10-year averaged, inflation-adjusted earnings. A long-term return predictor, not a timing signal.
- Buffett Indicator — total market cap / GDP. Whole-market valuation vs the economy.
- Equity Risk Premium (ERP) — expected excess return of stocks over bonds; falls when stocks are expensive.
- Yield curve / inversion — bond yield by maturity. Inversion (short > long) historically preceded recessions.
- Credit spreads / OAS — corporate-over-treasury yield difference. Widening = financial stress; tight = risk appetite / complacency.
- ISM / PMI — purchasing-manager survey indices; >50 expansion, <50 contraction.
- Breadth — how many stocks participate in the trend (e.g. % above 200-day MA). Narrow breadth with a rising index = warning.
- Sahm Rule — recession signal based on rising unemployment; lagging.
- Margin debt — debt taken to buy stocks. Coincident, not predictive.
- VIX — implied volatility of S&P 500 options, the “fear index”.
- AAII Bull-Bear — weekly retail sentiment survey; contrarian at extremes.
- Recession — loosely two quarters of GDP decline; in the US formally dated by NBER.
- Diversification — spreading capital across assets of different character to limit specific risk.
- Rebalancing — restoring target weights; trims what grew (euphoria), buys what cheapened (bear).
- Dollar-cost averaging (DCA) — investing a fixed amount at fixed intervals, regardless of price.
- Bear-market rally / dead cat bounce / bull trap — a counter-trend rally inside a bear that fades; a “false dawn”.
- Capitulation — phase of extreme panic and forced selling, usually near the bottom.
- Informational cascade / herding — imitating the crowd against your own analysis.
- Loss aversion — the pain of a loss (~2×) outweighs the joy of an equal gain.
- Minsky moment — the point where risk built up over a long calm suddenly materializes.
📖 Further reading/watching
- Goldman Sachs / Peter Oppenheimer, “Global Strategy Paper: Bear Market Anatomy”, 2025-04-08
- CFA Institute, “Bear Market Playbook”, 2025-07-22
- Hartford Funds / Ned Davis Research, “10 Things You Should Know About Bear Markets”, 2025
- Robert J. Shiller, Online Data (ie_data.xls) & Irrational Exuberance
- Reinhart & Rogoff, “The Aftermath of Financial Crises”, NBER WP 14656, 2009
- Kahneman & Tversky, “Prospect Theory”, Econometrica, 1979
- Howard Marks, Mastering the Market Cycle (2018); memos “Nobody Knows”, “The Anatomy of a Rally”, “Cockroaches in the Coal Mine”
- Charles Kindleberger & Robert Aliber, Manias, Panics, and Crashes
- Marcin Iwuć — Finansowa Forteca, marciniwuc.com (source author)
- Related: Crypto Market State 2026 · Fincept Terminal
Template: knowledge_note_info