Smart-Beta, Factor & Thematic ETFs

The single most-cited paper here is Ben-David, Franzoni, and Kim (2023): 'Competition for Attention in the ETF Space.' They studied roughly 1,000 ETF launches and found that thematic and specialized ETFs underperformed broad-market benchmarks by an average of about 3-4% per year in the five years AFTER launch. Smart-beta and single-factor ETFs decayed less but still meaningfully. Their explanation is structural: the ETFs that get launched are the ones that look attractive based on recent performance, and 'recent performance' is the worst possible signal for future performance. Vanguard's 2019 factor-investing white paper reaches a parallel conclusion from a different angle: factor premiums have shrunk meaningfully since their academic publication.

There's an even bigger risk hiding in thematic ETFs: when launch timing aligns with the theme's hype peak, the underlying stocks are mispriced UP at launch, the ETF buys at the peak, and the index methodology systematically reconstitutes at peaks (because reconstitution adds stocks with the most recent gains). The ARK Innovation ETF (ARKK) drawdown from late-2021 to 2022 is the canonical example -- the index design wasn't broken, but the launch and inflow timing turned a 5-year backtested 15%+ annualized into a live 5-year flat-to-negative outcome for the typical buyer.

None of this means factor investing is fraudulent. The classical factors (value, size, momentum, quality, low volatility) have decades of academic evidence and survive transaction costs at institutional scale. The retail problem is that the ETF wrapper bundles in 0.3-0.7% in expense ratios, 0.1-0.3% in tracking error, behavioral churn (investors exit during the inevitable 3-5-year drawdowns), and competition for the alpha (after the factor is well-known). The net for the retail investor is often worse than just owning VTI plus a fixed bond allocation.

Smart-beta promises a free lunch and delivers a discounted one. Backtest returns are systematically biased upward because the strategies that get published are the strategies that worked. Live returns reflect crowding, fees, tracking error, and behavioral timing -- net erosion of 2-5% per year on average. If you tilt toward a factor, do it knowingly, accept multi-year drawdowns, and consider whether a 0.03% S&P 500 index fund already gives you 90% of what the smart-beta fund is selling for 10x the cost.