Alpha

Alpha is the return a portfolio or fund earns above what its market exposure alone would predict. It is the part of performance that cannot be explained by simply moving with the index — the part attributable to skill, selection, or timing.

Formula

alpha = portfolio_return - (risk_free_rate + beta * (market_return - risk_free_rate))

This is alpha as defined by the Capital Asset Pricing Model. In plain terms: take the return the portfolio actually delivered, and subtract the return you would have expected given how much market risk it took on. What is left is alpha.

How to read it

• Alpha = 0 — the portfolio matched what its market exposure predicted. No skill demonstrated, no skill required.
• Alpha > 0 — the portfolio outperformed its benchmark on a risk-adjusted basis. A 2% alpha means the manager added 2 percentage points of return beyond market exposure.
• Alpha < 0 — the portfolio underperformed for the level of risk it carried. The active decisions subtracted value compared to a passive position with the same beta.

Alpha vs total return

A portfolio that returns 18% in a year the market returned 20% has positive total return but negative alpha — the manager captured less of the upside than the market exposure alone should have produced. This is the central reason alpha exists as a concept: raw returns are misleading, because they conflate market gravity with manager skill.

Limitations

• Hard to distinguish from luck. A few years of positive alpha can easily be random. Most academic work suggests you need 10+ years of data before alpha is statistically meaningful.
• Sensitive to benchmark choice. Alpha measured against the wrong benchmark is meaningless. A small-cap fund measured against the S&P 500 will show alpha that is really just the small-cap premium.
• Compresses over time. Most professional managers fail to deliver positive alpha after fees over long periods. Alpha is rare and fragile.