The Bubble You Can't Short | Rob Arnott on What You Can Do Instead

Episode Summary

Executive Summary: The episode argues that parts of the market—especially U.S. large-cap growth, AI leaders, and certain story stocks—fit a bubble definition based on implausible future growth and narrative-driven buying. The guest contrasts this with cheaper value, fundamental indexing, and non-U.S. markets, while also emphasizing that AI is a real technological revolution whose main economic benefits may accrue broadly, not to today’s infrastructure builders.

Main Topics: Defining bubbles in real time (Priority: 5/5): A bubble exists when current prices require implausible future growth and are driven by narrative rather than discounted cash flow discipline. The guest stresses that bubbles are asset-specific and warns against shorting them directly. Dot-com lessons applied to today (Priority: 5/5): The discussion compares today’s market leaders to the 2000 tech names, highlighting that disruptors can be disrupted, narratives can be wrong about speed and moat durability, and cap-weighted indices can mask long bear markets beneath index-level returns. AI as revolutionary but not necessarily profitable for builders (Priority: 5/5): AI is portrayed as a genuine, world-changing technology, but the companies spending heavily on AI capex may struggle to earn attractive returns on that spending. The guest distinguishes technological importance from investment attractiveness. Valuation, value investing, and fundamental indexing (Priority: 4/5): Research Affiliates’ RAFI approach is presented as a better way to own value by weighting firms by business size and rebalancing away from overextended names. Value is described as historically cheap and attractive versus U.S. growth. Reinventing growth and avoiding low-quality expensive stocks (Priority: 4/5): The guest argues growth should be defined by observed growth rates rather than price labels, and that portfolios should exclude expensive stocks with sluggish growth. Fast-growing, cheap companies outperformed the market historically. International and emerging markets opportunity (Priority: 4/5): Non-U.S. stocks, especially emerging markets value, are described as underappreciated and far cheaper than U.S. equities. The guest expects better long-run returns abroad versus U.S. large-cap growth. Index construction and buy-high/sell-low effects (Priority: 4/5): Traditional cap-weighted indexing is criticized for mechanically adding winners after they rise and selling losers after they fall. Alternative rules-based structures are presented as more sensible and potentially return-enhancing.

Key Arguments: A bubble can be identified when a stock’s price requires implausible, not impossible, growth assumptions and the marginal buyer is narrative-driven rather than valuation-driven. Bubbles do not need to crash immediately; they can rise much further and persist longer than expected, making shorting dangerous. Dot-com history shows that even dominant firms with moats can underperform for years or be disrupted; many 2000 leaders produced poor long-term returns. AI is transformative and likely beneficial to society broadly, but the companies funding massive AI capex may not convert that spending into profits. R&D spending has historically been positively associated with future stock success, while CapEx has been negatively associated with future stock performance. Value is unusually cheap relative to history, especially in the U.S., and international markets are priced at much lower multiples than U.S. equities. Fundamental indexing and the RAFI methodology can add value by reweighting away from overvalued giants and rebalancing toward cheaper fundamentals. A better growth index should select firms by actual growth and weight them by economic magnitude, avoiding expensive slow-growth names that are unlikely to justify price. The biggest winners from AI may be users and downstream businesses, not necessarily the current infrastructure suppliers or platform incumbents. Long-term portfolio construction should emphasize trimming expensive winners, averaging into cheap assets, and diversifying into value and non-U.S. exposures.

Data Points: Palantir market value: Over $500 billion - Used to illustrate a potential bubble relative to trailing revenue Palantir trailing 12-month revenue: $3 billion - Compared against its valuation to argue implied future growth is implausible Microsoft wait to beat S&P 500 after dot-com peak: 18 years - Example of how long it took a dot-com leader to outperform again NASDAQ decline after dot-com burst: Down 50% in 2 years; eventually down 80% - Shows the severity and persistence of tech bubble unwinding Median Russell 3000 stock return post-bubble: +20% - Illustrates that the average stock did not experience the same collapse as the cap-weighted index Russell 2000 Value return post-bubble: +53% - Shows value/small-cap outperformance while tech-heavy indices fell Magnificent Seven weight in S&P 500: 34% - Used to show concentration in U.S. large-cap growth Magnificent Seven weight in RAFI: 18% - RAFI downweights mega-cap concentration to economic footprint RAFI vs MSCI ACWI Value performance: +2.5% per year over 20 years - Global RAFI outperformance cited as evidence of rebalancing alpha RAFI tracking error: About 2% - Presented as modest relative to return advantage U.S. Shiller P/E: 40x - Argued to be historically expensive and near dot-com-era extremes Dot-com peak Shiller P/E: 44x - Reference point for current U.S. valuation extremes U.S. stocks in valuation history: Top 1% most expensive historically - Used to emphasize rich valuations Value cheapness ranking: Bottom 2% to 3% of historical cheapness - Indicates unusually attractive value conditions Emerging markets Shiller P/E: About 12x - Used to show much lower valuations than U.S. stocks Emerging markets value Shiller P/E: Single digits - Presented as especially compelling Expected return for emerging markets value: 10% annualized over 10 years - Forward-looking expectation from Asset Allocation Interactive Expected return for U.S. large-cap growth: 1.5% annualized over 10 years - Contrasted with EM value to support international/value overweight RACWI live outperformance vs S&P: +81 bps per year - Research Affiliates Cap Weighted Index performance claim over four years RACWI correlation to S&P: 99.96% - Indicates near-market exposure with small active tilts Portfolio overlap with S&P: 95% - Used to explain that small non-overlapping tilts are driving excess return RAFI historical t-statistic: Over 5 - Cited as unusually strong live-strategy evidence Growth index historical alpha: +4% per year incremental return - Claim for a new growth methodology based on observed growth Cheap and fast-growing quadrant performance: Beat the market by about 1% to 1.5% per year - Best-performing quadrant in a 55-year study Expensive and slow-growing quadrant performance: Underperformed the market by more than 2% per year - Worst-performing quadrant in the same study OpenAI revenue: $13 billion - Mentioned in the context of future spending commitments OpenAI stated spending/investment commitment: $1.4 trillion - Used to argue expectations are implausibly high AI-related spending vs revenues: Dozens of times larger - Characterization of current capex intensity versus realized revenues AI demographics research result: 1.3 billion people reached age 80 out of 100 billion ever lived - Example of using AI to quickly answer a research question

Pivotal Quotes: "You have to make implausible, not impossible, but implausible growth assumptions to justify today's price." — Rob Arnott: Core definition of a bubble "Never short-sell a bubble. Bubbles can go further and can last longer than you can possibly imagine." — Rob Arnott: Warning about trading against bubbles "If you go back historically and you sort companies based on R&D expenditure, that's positively correlated with subsequent success as a stock... CapEx tends to be negatively correlated with future stock market performance." — Rob Arnott: Argument on AI-era capital spending versus research spending

Implications: Investors should be wary of narrative-driven U.S. growth and AI enthusiasm, favor valuation discipline, and consider value, fundamental indexing, and international markets for better long-term expected returns.