Availability Cascade in Markets
Availability cascade in markets: how compounding information cascades amplify trends, create bubbles, and drive systematic mispricing.

Availability cascades drive bubbles + panics in markets. Recognising them is one of the most useful investment skills.
What is an availability cascade?
The Kuran-Sunstein framework.
Defined by Kuran + Sunstein (1999):
- 'Availability cascade' = self-reinforcing process where a belief gains more credibility through repeated public expression.
- Combines two effects: (a) Availability heuristic - things easier to recall feel more likely; (b) Social proof - what others believe affects our beliefs.
- Result: a narrative can become 'common knowledge' even when underlying evidence is weak.
In markets:
- Same dynamic creates bubbles + panics.
- Narratives drive price action; price action validates narratives.
- Feedback loop accelerates until the cascade hits a breaking point.
The 4-stage market cascade
How they unfold.
Stage 1 - Initial salient event:
- Real event happens (e.g. ChatGPT launches; novel coronavirus emerges; AI accelerates).
- Initial price reaction reflects new information.
- Could be reasonable response or initial overshoot.
Stage 2 - Media amplification:
- Financial media covers the story repeatedly.
- Each story adds 'updates' that feel new but mostly recycle initial claims.
- Repetition creates familiarity → familiarity feels like evidence.
Stage 3 - Social transmission:
- Discussion spreads beyond financial media (Twitter / Reddit / dinner parties).
- New investors enter who normally don't trade based on macro themes.
- Crowd psychology starts to dominate fundamentals.
Stage 4 - Cognitive availability ≠ statistical certainty:
- The narrative feels 'obvious' to everyone.
- People can't articulate WHY but everyone agrees.
- Counter-evidence is dismissed.
- Prices significantly diverge from fundamentals.
Eventually:
- Reality catches up.
- Cascade breaks - sharp reversal.
- Investors who recognised the cascade can profit; those who didn't get caught.
Examples - cascades that drove market moves
Historical pattern recognition.
Tech bubble (1995-2000):
- Initial story: 'Internet will change everything.'
- Media amplification: Wired, Fortune, business covers obsessed.
- Social transmission: 'My taxi driver is buying tech stocks.'
- Peak: NASDAQ 5,000 in March 2000.
- Reality: cascade broke; NASDAQ -75% by Oct 2002.
2008 housing bubble:
- Initial story: 'Housing always goes up.'
- Media + social: 'My neighbour made GBP 100,000 in a year on their house.'
- Bank lending standards collapsed (also enabled by cascade).
- Reality: cascade broke 2007-2008; global financial crisis.
2020-2021 COVID-era retail investing:
- Initial story: 'Tech + crypto are inevitable.'
- Media + Reddit (WallStreetBets) amplification.
- GameStop short squeeze + crypto run.
- Reality: cascade partially broke in 2022; large stocks recovered, crypto sustained losses.
2023-2024 AI narrative:
- Initial story: 'AI will transform productivity.'
- Media saturation; ChatGPT becomes daily news.
- Nvidia + AI-themed stocks dominate gains.
- Current question (2026): is this still a cascade or a fundamental reality?
Detection - cascade red flags
When you suspect one is forming.
- Story feels 'everywhere': media saturation; multiple parallel-coverage outlets.
- New entrants who don't normally invest are participating: 'My grandmother is asking about Bitcoin.'
- Story has clear simplifying narrative: complex reality reduced to memetic phrase.
- Counter-evidence is dismissed quickly: 'This time it's different' / 'They don't understand.'
- Price action diverges from fundamentals: P/E ratios, growth rates, addressable market all stretched.
- Social cohesion around the belief: skeptics are mocked or shunned.
- Multi-step value justifications: 'If A and B and C and D, then this stock is worth X' (conjunction fallacy embedded in cascade).
- Recent winners drive new participation: 'My friend made GBP 50,000 - I should buy too.'
Profiting from cascade recognition
Practical strategies.
Buying into cascades:
- Early cascade: undervalued; rational to participate.
- Middle cascade: fair value to mild premium; participate with caution.
- Late cascade: significant premium; reduce position or sell.
- Peak cascade: maximum premium; small short exposure may be rational.
Identifying cascade phases:
- Phase indicators: media coverage frequency, retail entry, fundamental divergence.
- Historical: average cascade peaks ~12-24 months from media saturation.
- 2026 examples: AI cascade probably middle-to-late; crypto cascade phase varies by coin.
Buying after cascades break:
- Post-cascade reversion creates good entry points for genuine long-term winners.
- Tech bubble: buying Amazon at 2002 lows was generational.
- 2022 tech reversion: created opportunities in fundamentally-strong companies.
- Requires patience + willingness to buy when everyone is selling.
Avoiding mistakes:
- Don't time the exact top - cascades can run longer than expected.
- Don't dismiss the underlying claim - just question the price.
- Don't sell entire position early - average out over the cascade.
Cascade vs paradigm shift
Distinguishing them.
Cascade:
- Price reflects sentiment more than fundamentals.
- Reverses when reality fails to match narrative.
- Long-term value typically below cascade-peak prices.
- Example: housing 2007.
Genuine paradigm shift:
- Price reflects fundamental new reality.
- Sustains over decades.
- Long-term value much higher than current price.
- Example: internet 1995 - cascade-overpriced but fundamental shift was real.
How to distinguish (real-time difficult):
- Reference class: 'What's the base rate of such claims actually being paradigm shifts vs cascades?'
- Independent reasoning: 'If I had no information from media, would this still feel transformative?'
- Time horizon: paradigms compound over decades; cascades reverse in years.
- Test: 'Imagine the cascade reversed - is the underlying value still defensible?'
Mitigation - protect yourself from cascades
Pre-commitment + reference class.
- Pre-commit to allocation rules: 'I won't exceed 10% in any single thematic sector.' Removes mid-cascade deliberation.
- Reference class forecasting: 'How often have similar cascades reversed?'
- Independent fundamental analysis: re-derive the value from first principles, not based on media coverage.
- Diversify information sources: read outside the cascade (e.g. competing perspectives, contrarian analysis).
- Decision journal: write down your reasoning + revisit later.
- Position sizing discipline: never risk losing your sleep on any single thematic bet.
- Tax-loss harvesting: when cascade breaks, harvest losses for tax optimisation while staying invested in fundamental winners.