Week 41 Day 4: Alpha: The Holy Grail Nobody Can Find

Alpha = your return minus the return explained by the risk you took. If VTI returned 10% and your portfolio (with the same risk) returned 12%, you generated 2% alpha. That means you added value beyond what the market gave for free. The problem: almost nobody consistently generates alpha. Over 15 years, 92% of large-cap fund managers earned NEGATIVE alpha (they underperformed VTI after fees).

The alpha scorecard (SPIVA, 2023): Over 1 year: approximately 60% of actively managed large-cap funds underperform the S&P 500. Over 5 years: approximately 75% underperform. Over 10 years: approximately 85% underperform. Over 15 years: approximately 92% underperform. Over 20 years: approximately 95% underperform. These are professional fund managers with armies of analysts, proprietary data, and Bloomberg terminals. They cannot consistently beat VTI. Why alpha is so rare: (1) Markets are competitive. Every buyer has a seller. For you to earn alpha, someone else must earn negative alpha. The market is a zero-sum game (before costs) and a negative-sum game (after costs). (2) Costs compound against you. A 1% management fee on a fund that otherwise earns market returns guarantees -1% alpha. After fees, even skilled managers have difficulty producing positive net alpha. (3) Alpha decays. When a strategy that generates alpha is discovered, money flows into it, reducing the opportunity. What worked in 2005 may not work in 2025. (4) Survivorship bias hides the losers. You see the funds that survived 20 years (the winners). You do not see the thousands that closed (the losers). The survivors' track records are biased upward. The conclusion: do not pay for alpha. Buy the market (VTI) at 0.03% fees and capture the market return. The 8% of managers who DO beat the market over 15 years cannot be reliably identified in advance.

Alpha, in the CAPM framework, is defined as alpha_i = R_i - [R_f + beta_i * (R_m - R_f)], where R_i is the realized return, R_f is the risk-free rate, beta_i is the portfolio beta, and R_m is the market return. Under a multi-factor model (Fama-French, Carhart), alpha is the intercept after controlling for market, size, value, and momentum factors. Jensen (1968) introduced 'Jensen's alpha' as the measure of fund manager skill and found that the average alpha across mutual funds was approximately -1.1% per year -- fund managers as a group destroyed value after fees. Subsequent research has consistently confirmed this finding: Fama and French (2010) used bootstrap analysis to show that the cross-section of fund alphas is consistent with pure luck (the distribution of alphas matches what you would expect if ALL managers had zero skill, with some randomly outperforming and some underperforming). Barras, Scaillet, and Wernart (2010) estimated that approximately 75% of funds have zero alpha (no skill), approximately 24% have negative alpha (skill-destructive, possibly due to fees and overtrading), and only approximately 0.6% have genuine positive alpha. The 0.6% cannot be reliably identified in advance: Carhart (1997) showed that the correlation of fund performance between adjacent 5-year periods is approximately zero -- past performance does not predict future performance. The aggregate implication: paying for alpha (through active management fees) is a losing proposition in expectation. The expected alpha of a randomly selected active fund is approximately -1% per year (the average fee drag). Index funds guarantee alpha of approximately 0% (before minimal fees of 0.03%), which exceeds the expected alpha of active management by approximately 1%.