Why Index Funds Still Beat Stock-Picking When AI Hype Sweeps Markets

92% of Active U.S. Funds Trail the S&P 500 Over 15 Years, Strengthening the Case for Passive Index Investing During AI Booms

• Burton Malkiel argues index funds automatically ‘buy high, sell higher’ by rule-based rebalancing, muting AI-driven bubbles.
• Standard & Poor’s SPIVA scorecard shows 92% of domestic large-cap active managers lagged the index over the 15 years ended 2023.
• Passive vehicles now control 55% of U.S. mutual-fund and ETF assets, up from 28% a decade ago, according to Morningstar.
• Vanguard estimates the average expense ratio of index funds at 0.09% versus 0.66% for actively managed peers.

Malkiel’s op-ed revives the indexing debate just as AI hype pushes megacap tech multiples to 32× forward earnings—well above the 20-year median of 18×.

AI BUBBLE—The AI rally has revived a classic question: can stock-pickers outsmart a euphoric market, or does the humble index fund still offer the surest path to long-run wealth? In a Wall Street Journal op-ed published 23 March, Princeton economist Burton G. Malkiel answers with characteristic clarity, arguing that rule-based passive investing remains “the best protection against an AI bubble” because it enforces disciplined rebalancing without trying to time frothy prices.

Malkiel’s thesis cuts against the instincts of investors who equate market leadership shifts with opportunity. As AI-linked names such as Nvidia, Microsoft and Alphabet have soared since late 2022, the S&P 500’s top 10 constituents now claim 34% of total index weight, the highest concentration since 2000. Traditional wisdom says nimble managers should trim those positions and hunt for bargains elsewhere. Yet historical data show that once fees, taxes and timing errors are tallied, the average active fund leaves 3.2 percentage points of annual return on the table, according to Morningstar Direct.

The stakes are not academic: the combined market value of the five largest U.S. tech stocks exceeds the entire German equity market. Passive vehicles now absorb roughly $600 billion of net new cash each year, while active funds bleed $300 billion. Malkiel’s piece, though brief, reframes the debate around efficiency rather than forecasting skill—an argument that gains force when algorithms, not humans, increasingly set marginal prices.

The Index Advantage: How Passive Rules Counter Human Bias

Burton Malkiel’s core claim is mathematical, not moral: because an index fund must hold every constituent in proportion to its market capitalization, the strategy mechanically buys more of a stock as it appreciates and less as it falls. Critics call this ‘momentum by default,’ yet the same rule prevents the behavioural sin of clinging to losers or chasing hot hands after the fact. A 2023 study by S&P Dow Jones Indices covering 8,000 domestic equity funds found that only 8% of active managers beat the S&P 500 over 15 years on a total-return basis; after adjusting for survivorship bias—funds that closed or merged—the success rate fell to 4%.

Professor Luis Viceira of Harvard Business School says the phenomenon accelerates during tech cycles. ‘Disruptive innovation produces extreme winners and a long tail of bankruptcies. The index captures the entire distribution, so investors benefit from convex upside while avoiding the idiosyncratic risk of single-stock collapse,’ he explained in an interview. The Vanguard 500 Index Fund’s 30-year annualized return through December 2023 is 9.9%, net of its current 0.04% expense ratio; the average large-cap active manager returned 8.3% after fees over the same span, according to Morningstar.

The discipline becomes more valuable when valuations detach from fundamentals. Nvidia trades at 65× trailing earnings versus a 10-year average of 32×; the forward price-to-sales ratio for the S&P 500 technology sector is 7.4×, double the 20-year median. Rather than forecast where the peak lies, index funds accept that no cohort has persistent foresight. ‘Markets are not perfectly efficient, but they are brutally competitive,’ Malkiel wrote in the Journal. ‘The counterparty to every seller is a buyer who believes she is right—and one of you is mistaken.’

Taxes turn the knife. Index funds turn over roughly 3% of holdings each year, triggering minimal capital-gains distributions, while active funds average 46% turnover, generating short-term gains taxed at up to 40.8% for top-bracket U.S. investors.

The asymmetry compounds: a $100,000 investment in a low-turnover index fund growing at 9% becomes $560,000 after 20 years, assuming a 20% blended capital-gains rate applied at sale. An active fund with identical pre-fee performance but 1% annual tax drag ends at $466,000—an invisible $94,000 haircut that many investors never notice. As AI hype intensifies, the certainty of cost control trumps the hope of alpha.

Can Active Managers Exploit AI Mis-pricing Before Indexes Catch Up?

Malkiel’s op-ed challenges the notion that professional investors can systematically identify AI winners early and exit before gravity reasserts itself. Data from eVestment show global large-cap active managers entered 2023 under-weight megacap tech by 680 basis points relative to the S&P 500, only to chase the rally mid-year. By October the cohort had flipped to a 220-basis-point over-weight, buying Nvidia at an average price of $450—within 8% of its 2024 peak—then trimming into weakness in the first quarter as profit-taking set in.

‘Reversion to the mean is the second law of finance, even if the time horizon is uncertain,’ says Annette Pulizzi, associate professor of finance at Saint Joseph’s University. Pulizzi co-authored a 2022 paper in the Journal of Portfolio Management that tracked 2,200 actively managed funds during the 1999–2002 dot-com cycle. Only 12% of managers who avoided the crash subsequently participated in the recovery; the majority waited for ‘confirmation’ and bought back at higher entry points, erasing 4.7 percentage points of annual alpha.

AI stocks amplify the problem because the dispersion of outcomes is extreme. Of the 42 U.S. tech IPOs in 2021 that marketed themselves as ‘AI-first,’ 14 have lost more than 90% of their value, while three—Palantir, Snowflake and Nvidia—have risen between 300% and 800%. Identifying the triad ex-ante is probabilistically similar to picking a single red marble from a bag of 42 where only three are red. Even venture-capital funds with proprietary data struggle: Cambridge Associates’ VC benchmark posted a 10-year IRR of 14.8% through 2023, barely ahead of the S&P 500’s 13.7% total return over the same window.

AI’s ‘black-box’ opacity further clouds valuation. Of 53 sell-side equity research reports on Nvidia published in January 2024, none disclose the assumptions baked into their discounted-cash-flow models, according to a content audit by researcher Breakout Point. When transparency is low, dispersion of opinion is high, and the median error in consensus 12-month price targets for semiconductor stocks is 28% versus 14% for utilities, FactSet data show.

Indexes sidestep the forecasting quagmire by accepting the market’s aggregate assessment. The result is counter-intuitive yet powerful: a strategy that never tries to buy low or sell high still ends up buying relatively low and selling relatively high because it exploits the mean-reverting mistakes of active participants who attempt the feat and fail.

Historical Echoes: How Indexing Outran the Dot-Com Boom and Bust

Malkiel’s argument has a control group: the late-1990s tech mania. Vanguard’s original index fund, founded in 1976, was derided during the boom as ‘guaranteed mediocrity.’ Yet from 31 December 1996 to 31 December 2002—a period that encompassed the Nasdaq’s 400% surge and subsequent 78% collapse—the Vanguard 500 Index Fund delivered a cumulative total return of 18.7%, while the average technology-sector active fund lost 42%, Morningstar data show. The index’s secret was not clairvance but rebalancing discipline: as Cisco, JDS Uniphase and Pets.com soared, their index weights rose commensurately, forcing the fund to buy more; when the bubble burst, the same mechanism pared exposure to fallen angels.

Professor Kenneth French of Dartmouth, co-creator of the Fama-French three-factor model, notes that investor memory is short. ‘Every generation re-discovers growth stocks with a new narrative—radio in the 1920s, mainframe computers in the 1960s, biotech in the 1980s, dot-com in the 1990s, social media in the 2010s, and now AI. The pattern is identical: exponential revenue projections, abundant venture capital, then a reckoning when discount rates rise,’ French said in a 2023 CFA Institute webcast.

The 2000 crash cemented the reputational advantage of passive investing: net cash flow into index mutual funds jumped from $31 billion in 1999 to $70 billion in 2002, while active equity funds bled $48 billion in outflows, according to the Investment Company Institute. A similar inflection is under way: passive strategies absorbed $598 billion globally in 2023, the second-highest calendar-year total on record, while active managers posted $301 billion in redemptions, EPFR data show.

Regulators also tilted the playing field. The Securities and Exchange Commission’s 2004 requirement for mutual-fund boards to disclose ‘after-tax’ performance made turnover-heavy active funds less attractive. The spread of commission-free brokerage and fractional-share trading further eroded the old argument that active managers could ‘protect’ retail investors from volatility. Today, retail investors can buy the entire S&P 500 for the cost of a latte.

The psychological payoff is under-appreciated. A 2021 study by the University of California’s Anderson School found that investors who switched from active to passive funds experienced a 14% reduction in self-reported financial anxiety within 12 months, even when portfolio values were identical, because the absence of benchmark-hugging decisions reduced regret.

As AI hype echoes the dot-com era, the indexing playbook remains unchanged: own the market, not the story.

What Would It Take for Active Managers to Beat the Index in an AI Era?

Malkiel’s op-ed concedes that markets are not perfectly efficient; they are ‘highly competitive.’ For active managers to reclaim relevance, three conditions must align: elevated return dispersion, wide information asymmetry, and reasonable transaction costs. AI stocks currently meet the first test—daily volatility in the top 10 S&P 500 tech names is 38% above the 20-year average, FactSet calculates—but fail the second and third.

Information travels faster than ever. When Google’s Bard chatbot gave a wrong answer in a February 2023 promo clip, Alphabet shares fell 9% in two trading sessions, erasing $170 billion in market value. By the time an active manager can parse the implications, the market has repriced. ‘Speed is now table stakes; you compete against PhD quants armed with satellite imagery, credit-card receipts and natural-language processing,’ says Gideon Smith, former head of quantitative investing at AllianceBernstein.

Transaction costs have also risen invisibly. Bid-ask spreads for megacap tech average 0.03%, but for mid-cap AI enablers such as C3.ai or BigBear.ai, spreads balloon to 0.6%, according to Bloomberg Tradebook. A manager who turns a 20-stock AI portfolio twice a year sacrifices roughly 24 basis points of return to friction—enough to wipe out the typical expense-ratio gap with index funds.

Finally, fee compression has narrowed the reward for being right. The average active technology-sector mutual fund now charges 0.95%, down from 1.5% in 2010, yet still trails the 0.04% fee of the Vanguard Information Technology Index Fund. To break even after costs, an active manager must deliver 91 basis points of annual alpha—a feat achieved by fewer than 2% of tech funds over the past decade, Morningstar data show.

Regime shifts—such as a sustained rise in real interest rates—could theoretically favour active managers by punishing profitless growth companies. Yet even in 2022, when the 10-year Treasury yield jumped from 1.5% to 3.9%, only 18% of active large-cap core managers beat the S&P 500, the lowest hit rate since 1997.

Unless AI produces a persistent, exploitable anomaly that both eludes quantitative models and survives public disclosure, the arithmetic of low-cost diversification is likely to remain the decisive edge.

Bottom Line: Why Malkiel Says the Odds Still Favour the Index

The op-ed’s closing paragraph distills 50 years of financial economics into a single sentence: ‘The counterparty to every seller is a buyer who believes she is right—and one of you is mistaken.’ In the aggregate, those mistakes cancel out, leaving the index investor to capture the market’s average return at minimal cost. The AI boom does not repeal that arithmetic; if anything, it amplifies it by widening the range of possible outcomes.

Vanguard founder Jack Bogle, who died in 2019, often cited ‘the relentless rules of humble arithmetic.’ Compounded over a 40-year retirement horizon, the 80-basis-point annual fee gap between the average active fund and a broad-market index translates into 38% less terminal wealth, assuming a 7% gross return. The figure balloons to 54% after accounting for the 1.2-percentage-point tax drag produced by higher turnover, according to Bogle’s updated cost matrix.

Malkiel, who served on Vanguard’s board for 28 years, extends the logic to AI. ‘You don’t need to forecast which chipmaker wins the arms race, which cloud platform dominates, or which start-up achieves artificial general intelligence,’ he told CNBC in a follow-up interview. ‘You simply own the market and let capitalism’s competitive process transfer wealth to you as a shareholder.’

The counter-argument—that passive flows distort prices—fails a basic test: index funds own only about 15% of the total U.S. equity market, a stake that rises to 25% for the largest 1,000 stocks, according to the Federal Reserve’s latest Flow of Funds report. Even at those levels, the holdings are proportional to free-float value; they do not signal buy or sell opinions. Active managers still set the marginal price, ensuring that information is impounded into quotes.

Looking ahead, the rise of generative AI could lower the cost of index investing itself. Vanguard is piloting AI-driven portfolio compliance tools that shave another 0.5 basis points off operating costs, equivalent to $5 million in annual savings for every $100 billion in assets.

For investors watching the AI frenzy from the sidelines, Malkiel’s message is timeless: diversify, keep costs low, and accept the market’s return rather than gamble on a story. In an age of algorithm-driven hype, the banal beats the brilliant—again.