Michael Mauboussin - Pattern Recognition and Public Markets (EP.370)

Episode Summary

Executive Summary: Michael Mobison argues that pattern recognition is powerful only in stable, linear environments with reliable base rates, and dangerous in complex domains like markets where intuition often becomes storytelling. The conversation then shifts to public markets, where shrinking listings, extreme return skew, and the dominance of a tiny fraction of superstar companies reshape how investors should think about indexing, manager selection, and long-term compounding.

Main Topics: Pattern recognition vs. expertise (Priority: 5/5): Mobison distinguishes real expertise from mere experience: intuition is useful when causal relationships are stable and predictive, but it breaks down in complex, reflexive systems like markets. When intuition works and when it fails (Priority: 5/5): Using chess as the clean example and macro forecasting as the weak one, he explains that pattern recognition is most reliable in structured, repeatable settings and least reliable when the environment changes. Decision aids in investing (Priority: 4/5): He argues fundamental investors should use more quantitative decision aids such as base rates and structured position sizing, while quants should demand more economic narrative behind factors. Public market returns and expected returns (Priority: 4/5): Mobison notes that public market expected returns improved meaningfully from 2022 to 2024, making the asset class more attractive than during the prior low-return period. The shrinking number of public companies (Priority: 4/5): He discusses the decline in listed companies from the 1990s peak to today, driven by fewer IPOs and ongoing M&A and buyouts, with much of the decline concentrated in microcaps. Power laws and wealth creation (Priority: 5/5): Bessenbinder’s research shows that a tiny fraction of public companies create nearly all wealth, implying that indexing and concentrated selection are rational responses to extreme outcome skew. Compounding, drawdowns, and staying power (Priority: 4/5): Great compounders often suffer severe interim drawdowns, so holding the eventual winners requires patience, conviction, and the ability to endure long periods of underperformance.

Key Arguments: Pattern recognition is best viewed as intuitive expertise: useful only when the environment is stable, relationships are linear, and causal patterns persist. Experience alone does not make someone an expert; an expert is someone whose predictive model actually works. Markets are complex adaptive systems, so expert forecasts often fail; simple extrapolation can match or beat them. Fundamental investors should use base rates and more structured decision aids, especially for sizing and manager selection. Public markets still offer attractive expected returns relative to the low-return environment of 2022, though not dramatically above long-term norms. The decline in public company counts is largely due to fewer IPOs and continued M&A/buyout exits, with the largest impact in microcaps. Public market wealth creation is extraordinarily skewed: a very small number of firms drive nearly all value creation, which supports both broad indexing and selective concentrated investing. Long-term winners often endure massive drawdowns, so capturing compounding requires patience and resilience. Large companies, especially leading tech firms, may sustain unusually high returns on capital due to intangible assets, scale, and network effects. Investors should focus on strengths and use mental models and cross-disciplinary learning to improve decision-making.

Data Points: CounterPoint Global AUM: $70 billion - Describes Michael Mobison’s firm 2022 expected nominal equity return: 5.8% - From Aswath Damodaran-based estimates for public markets at the start of 2022 2022 expected real equity return: 3.2% - Nominal expected return minus inflation expectations of a little over 2.5% Historical U.S. real equity return: 6% to 7% - Used as the long-run comparison for expected returns 2024 expected nominal equity return: 8.5% - Expected returns at the start of 2024 U.S. public companies in mid-1970s: About 4,800 - Historical count around the launch of the Wilshire 5000 Peak U.S. public companies: A little over 7,300 in 1996 - Peak number of listed companies Current U.S. public companies: Around 4,200 - Approximate 2023 count IPO average, 1976-2000: About 280 per year - Historical average IPO pace IPO average since 2001: About 120 per year - Lower IPO pace in the 21st century U.S. IPOs in 2022: 38 - Very weak IPO market U.S. IPOs in 2023: 54 - Still depressed versus historical norms Buyouts as share of delistings: About 20% - In the last 10 to 15 years, buyouts accounted for about one-fifth of public-company delistings Total public companies studied by Bessenbinder: Little over 28,000 - U.S. public companies since the 1920s Companies earning less than T-bills: Just under 60% - Monthly return result from Bessenbinder’s research Wealth destroyed by underperformers: $9 trillion - Aggregate wealth destruction from companies that underperformed Treasury bills Wealth created by surviving/value-creating firms: $64 trillion - Aggregate wealth creation by companies that created value Total net wealth creation: $55 trillion - Wealth creation minus destruction through 2022 Wealth concentration: Top 2% created $50 trillion of $55 trillion - Shows extreme right-tail concentration of public-market outcomes Venture capital deal sample: 30,000 deals - Referenced to show similar skewness in private markets Venture deals not returning capital: 60% - Share of VC deals that did not earn back the invested dollar Average venture holding period: Five to six years - Used to compare VC time horizons with public-market dynamics Public equity portfolio example at Berkshire Hathaway: Just under half in one stock (Apple) - Illustrates concentration risk in owning superstar winners Typical corporate hurdle rate: Roughly 15% - Companies often use a hurdle rate near 15% regardless of lower theoretical cost of capital Great compounder drawdowns: 75%+ common; some 80%-85% - Bessenbinder finding on interim drawdowns among top wealth creators

Pivotal Quotes: "all experts have experience, but not all experienced people are experts" — Michael Mobison: Explaining the difference between accumulated experience and true predictive expertise "people price companies and they don't value companies" — Aswath Damodaran (quoted by Mobison): Used to criticize overreliance on multiples without understanding economics "the top 2% of that 28,000 created 50 trillion of the 55 trillion in total" — Michael Mobison: Summarizing the extreme skew in public-market wealth creation

Implications: Investors should be far more selective about when to trust intuition, use more structured decision aids, and accept that public-market returns are driven by a tiny set of winners. Patience, concentration, and base-rate thinking matter more than storytelling.