You Can’t Sell Origami

Even perfect foresight cannot eliminate severe drawdowns, so investors should repair their mental models rather than abandon a sound investment thesis after a difficult period.

• The endowment effect makes investors overvalue what they own, including their investment models; a 1990 Cornell mug experiment found only four to eight trades across five rounds versus roughly eighteen predicted, with owners demanding 1.8 to 3.8 times buyers’ valuations.
• A hypothetical God Portfolio selected the 25 best-performing U.S. stocks in hindsight from the 500 largest companies and held them for five years; across 307 overlapping cohorts from December 1995 to June 2026, it earned an average 348% total return but suffered an average 31% maximum drawdown.
• Longer holding periods and greater concentration increased both returns and pain: the 10-stock, 20-year portfolio returned an average 6,475%, equivalent to compounding at 22.5% annually, but experienced a 54% average drawdown.
• The best possible 25-stock portfolio formed on June 30, 2021 returned 596% through June 2026, with Micron gaining nearly 1,300%; nevertheless, the portfolio fell 31% while its source index fell 25%.
• AQR’s U.S. Quality Minus Junk series turned $1 into $10.65 over 69 years, with 99.2% of rolling ten-year periods positive, but by June 30, 2026 it was 36% below its peak after its deepest drawdown, lasting 32 months and including four consecutive losing years.
• The recent quality slump invalidates the idea that quality reliably protects over short horizons, but it does not by itself disprove the deeper claim that durable competitive advantages and persistent high returns on invested capital can be underpriced.

Takeaway: Investors should discard the belief that quality works consistently or cushions every short-term shock, while preserving and reassessing the broader thesis around durable competitive advantages; severe drawdowns are compatible with strong long-term returns, but the durability of individual moats must be examined in light of AI and the capital-expenditure cycle.