Week 41 Day 5: The Cost of Complexity: Why Simple Portfolios Win

The Bogleheads three-fund portfolio: VTI (U.S. stocks) + VXUS (international stocks) + BND (bonds). Total expense ratio: approximately 0.05%. Performance: matches or beats the vast majority of professional managers over 10+ year horizons. It takes 30 minutes to set up, 30 minutes per year to rebalance, and zero minutes worrying about stock picks.

Simple vs. complex -- the evidence: (1) The three-fund portfolio vs. institutional endowments. Over 2003-2023, a simple 60/30/10 VTI/VXUS/BND portfolio outperformed the MEDIAN university endowment (which uses private equity, hedge funds, venture capital, commodities, and pays millions in consultant fees). The simple portfolio costs 0.05%. The endowment costs 2-3%. (2) Warren Buffett's bet. In 2008, Buffett bet $1 million that an S&P 500 index fund would beat a selection of hedge funds over 10 years. The index fund won handily: +125% vs. +36% for the hedge fund portfolio. (3) The complexity tax. Each additional fund in a portfolio adds: decision fatigue (which one to buy/sell), rebalancing complexity, tax complexity (tracking cost basis across funds), and the temptation to tinker (tactical shifts that usually subtract value). Approaches ranked by complexity and expected outcome: Tier 1 (simplest, excellent): VTI + BND (two funds). Tier 2 (simple, excellent): VTI + VXUS + BND (three funds). Tier 3 (moderate, excellent): VTI + VXUS + BND + VTIP + SCHD (five funds). Tier 4 (complex, might be slightly better OR worse): add factor tilts, alternatives, tactical shifts. Each tier above Tier 2 adds marginal improvement of 0.1-0.3% but substantial complexity. The effort-to-return ratio of Tiers 1-2 is far higher than Tiers 3-4. For most investors, Tier 2 is optimal.

The superiority of simple portfolios has multiple theoretical explanations: (1) Estimation error (DeMiguel et al., 2009). Complex strategies require more input parameters (expected returns, covariances, factor loadings), each estimated with error. The errors compound, producing portfolios that are optimal for the estimated parameters but suboptimal for the TRUE parameters. Simple strategies (1/N, market-cap-weighted) require zero parameter estimation and are therefore robust to estimation error. (2) Overfitting (Harvey, Liu, and Zhu, 2016). Complex strategies can be backtested on historical data and appear to generate alpha, but the 'alpha' is often a statistical artifact of searching many strategies and selecting the best-performing ones post-hoc. Adjusting for multiple testing, most published 'factors' and 'strategies' have likely zero true alpha. (3) Implementation leakage. Complex strategies have higher trading costs (more frequent rebalancing), higher tax costs (more taxable events), and higher behavioral costs (more decisions where biases can intervene). These costs can easily consume the small theoretical alpha that complexity provides. Sharpe (1991) proved the arithmetic of active management: the aggregate return of all active investors equals the aggregate return of all passive investors (both hold the market), minus the higher fees of active investors. Therefore, the AVERAGE active investor must underperform the AVERAGE passive investor by exactly the fee differential. Complexity does not create returns; it creates costs. The passive approach captures the market return efficiently (0.03% cost), and any complexity-driven alpha must exceed its own costs to add value -- a bar that 90%+ of complex strategies fail to clear.