What Is Alpha in Finance? Definition and Examples

Search Fintrev

Search our guides, reviews and glossary.

Alpha (α) is the part of an investment’s return that cannot be explained by the market or by the risk the investment took. In plain terms, alpha in finance is the excess return over a benchmark, usually adjusted for risk. An alpha of +2 means the investment beat its risk-adjusted benchmark by 2 percentage points over the period measured. An alpha of −2 means it fell short by the same amount, and an alpha of zero means it earned what its benchmark and risk level would predict.

Investors often call alpha a manager’s “edge,” because it is meant to isolate the value a portfolio manager adds or subtracts through stock selection, timing or strategy, separate from simply riding the market up or down.

How Alpha Is Calculated

The word “alpha” covers more than one calculation, so the first question with any alpha figure is which method produced it.

Simple Alpha (Excess Return)

The most basic version compares total returns:

Alpha = Investment return − Benchmark return

If a large-cap U.S. stock fund returned 12% in a year and its benchmark, such as the S&P 500, returned 10%, simple alpha is +2%. This version is easy to read, and fund ratings often show it as a single number like +3.0 or −5.0, meaning 3% better or 5% worse than the benchmark.

Its weakness is that it ignores risk. A fund that holds more volatile stocks than its benchmark will tend to beat the benchmark in rising markets simply because it is more exposed to the market. Simple alpha would credit that to skill.

Jensen’s Alpha (Risk-Adjusted Alpha)

Economist Michael Jensen introduced a risk-adjusted version in 1968 while evaluating mutual fund managers. Jensen’s alpha uses the CAPM to estimate what a portfolio should have earned given its beta, then measures the difference:

Alpha = Rp − [Rf + β × (Rm − Rf)]

Where:

• Rp = the portfolio’s actual return
• Rf = the risk-free rate (typically a short-term Treasury yield)
• β = the portfolio’s beta relative to the market benchmark
• Rm = the market benchmark’s return
• Rm − Rf = the market risk premium

The bracketed term is the CAPM expected return. Anything the portfolio earns above it is alpha; anything below it is negative alpha.

Alpha as a Regression Intercept

In practice, analysts often estimate alpha by regressing a portfolio’s excess returns (return minus the risk-free rate) against the market’s excess returns over many periods, such as 36 or 60 monthly observations:

Rp − Rf = α + β(Rm − Rf) + ε

Beta is the slope of that line and alpha is the intercept. This is where the Greek letter comes from: α and β are the standard symbols for the intercept and slope in a linear regression. Because the regression often uses monthly data, the monthly alpha is commonly annualized so it can be compared with annual returns.

Worked Example: Positive Alpha With Lower Returns

Hypothetical example for illustration only. Assume a one-year period in which:

• The risk-free rate is 3%
• The market benchmark returns 10%, so the market risk premium is 7%

Fund A is a defensive equity fund with a beta of 0.8. It returns 9%.

• Simple alpha: 9% − 10% = −1%
• CAPM expected return: 3% + 0.8 × 7% = 8.6%
• Jensen’s alpha: 9% − 8.6% = +0.4%

Fund A lagged the market, yet it has positive risk-adjusted alpha. With only 80% of the market’s sensitivity, the CAPM expected it to earn 8.6%, and it earned more.

Fund B is an aggressive growth fund with a beta of 1.3. It returns 13%.

• Simple alpha: 13% − 10% = +3%
• CAPM expected return: 3% + 1.3 × 7% = 12.1%
• Jensen’s alpha: 13% − 12.1% = +0.9%

Fund B looks like a big winner on simple alpha, but most of its outperformance came from taking extra market risk. Only 0.9 percentage points remain after the risk adjustment.

Interpreting Alpha

Positive alpha means the investment beat its benchmark (or its CAPM-predicted return) over the period, zero alpha means it tracked it exactly and negative alpha means it underperformed. For a portfolio meant to match or beat an index, negative alpha means it is falling short of that goal.

A “good” alpha depends on the investor’s goals and risk tolerance, but the basic threshold is anything above zero after adjusting for risk and fees. A few other points shape the reading:

• Units matter. Alpha is usually expressed in percentage points per period. An alpha of 0.5 per month is very different from 0.5 per year.
• Persistence matters more than one period. A single year of +4% can come from luck. Alpha that holds across several years and market conditions carries more information about skill.
• Statistical confidence matters. A regression-based alpha comes with a standard error. A small positive alpha with a wide margin of error may not be distinguishable from zero.

Alpha vs. Beta

Alpha and beta answer different questions, and they are reported together because each is incomplete alone.

Beta measures how sensitive an investment is to movements in the overall market, its systematic risk. A beta of 1.0 moves in line with the market, above 1.0 amplifies market moves and below 1.0 dampens them.

Alpha measures the return left over after accounting for that market exposure.

A useful way to separate them: beta returns can be bought cheaply through a low-cost index ETF, while alpha is what active management claims to deliver on top, and what its fees pay for. “Smart beta” funds, which track rules-based subsets of an index such as dividend growers or low-volatility stocks, sit between the two. Their extra return over a broad index may reflect a systematic tilt rather than manager skill.

Alpha and Other Risk Measures

Alpha is one of five statistics commonly used in modern portfolio theory to describe an investment’s risk-return profile. The others are:

• Beta: sensitivity to the benchmark.
• Standard deviation: how widely returns vary around their average, a measure of total volatility.
• R-squared: the share of an investment’s movements explained by the benchmark, on a scale of 0 to 100. A high R-squared means the benchmark fits the investment well, so its alpha and beta are more meaningful. A low R-squared suggests the benchmark is a poor match and the alpha figure may be misleading.
• Sharpe ratio: excess return over the risk-free rate per unit of total volatility.

The Sharpe ratio and Jensen’s alpha both adjust for risk but in different ways. The Sharpe ratio uses total volatility and needs no benchmark. Jensen’s alpha uses only market (systematic) risk and depends on the benchmark chosen. A concentrated portfolio can show positive alpha while carrying a lot of stock-specific risk that alpha does not penalize but the Sharpe ratio does.

Alpha is also different from plain return measures like return on investment (ROI) or CAGR. Those describe how much an investment earned. Alpha describes how much it earned relative to what it should have earned given its benchmark and risk.

How Alpha Is Used

Evaluating and ranking funds. Alpha is a standard figure in mutual fund and ETF research, used to compare active managers against their benchmarks. Portfolio managers pursue it by selecting securities they believe are mispriced, while diversifying to keep unsystematic (company-specific) risk from dominating results.

Hedge funds. In the hedge fund world, alpha is close to the whole product. Many strategies aim to deliver returns that do not depend on the market’s direction, for example by hedging out market exposure so that what remains is mostly alpha. Fees in this space are often justified by claims of alpha, which makes measuring it correctly especially important.

Quantitative research. At quant firms, “alpha” has a second, related meaning: a signal or model that predicts future returns. “Alpha research” is the work of finding, testing and combining such signals, for example a pattern in earnings revisions or price momentum that historically preceded outperformance. A signal “has alpha” if it is expected to produce returns not explained by known risk factors. Quant researchers also watch for “alpha decay,” where a signal’s predictive power fades as more traders exploit it.

Weighing fees. Alpha should be compared with what it costs. Hypothetical example: an advisor charges 1% of portfolio value per year and generates 0.75% of alpha before fees. The advisor did improve the portfolio’s performance, but the client is 0.25 percentage points worse off than in a fee-free benchmark. By the same logic, an alpha of zero before fees is a net loss once any fee is charged. Whether a published alpha is gross or net of fees changes its meaning entirely.

Alpha and the Efficient Market Hypothesis

The efficient market hypothesis (EMH) holds that market prices reflect all available information, so securities are generally priced correctly. Under the strong reading of the EMH, consistently earning alpha is impossible: mispricings either don’t exist or are corrected too quickly to exploit, and anomalies are rare.

The evidence on active funds is consistent with alpha being hard to achieve. One frequently cited comparison of active mutual funds against their passive benchmarks found that fewer than 10% earned positive alpha over periods of 10 years or longer, and the share drops further after taxes and fees. Results like this are a major reason many investors have moved to index funds and low-cost robo-advisors, preferring to match the market rather than pay to try to beat it.

There is also a theoretical objection to taking alpha at face value. Because beta risk can be managed through diversification and hedging, some argue that measured alpha is often compensation for risks the model does not capture. If a fund earns extra return by holding small, illiquid or financially weaker companies, the CAPM will count that as alpha, while multi-factor models that add size, value and other factors may attribute much of it to those exposures instead.

Limitations of Alpha

The benchmark determines the answer. Alpha is only as meaningful as the benchmark it is measured against. Comparing a convertible bond fund with a plain investment-grade bond index can make it look like it produced large alpha, when convertibles carry stock-like risk the index does not. Comparing a dividend-growth equity fund with the S&P 500 may misattribute returns to skill when they really come from a tilt toward a particular slice of the market. Where no suitable index exists, analysts sometimes construct a custom or simulated benchmark, which adds its own assumptions.

Alphas are not comparable across asset classes. Simple alpha is calculated against a benchmark in the same asset category. An equity fund’s alpha against a stock index and a bond fund’s alpha against a bond index answer different questions, so a +2 for one is not better or worse than a +1 for the other.

Time period and market cycle. Alpha measured over a strong bull market can look very different from alpha over a full cycle. Short windows are especially sensitive to a few large moves, and different market cycles affect alphas across asset classes differently.

Backward-looking. Alpha describes past performance. A manager’s historical alpha does not guarantee future alpha, and strategies that worked can stop working once widely copied.

Survivorship and selection. Funds that performed poorly are often closed or merged, so databases of surviving funds can make average historical alpha look better than investors actually experienced.

FAQs

The Bottom Line

Alpha matters most to investors weighing active funds, hedge funds or advisors against cheaper index options. A single alpha figure says little until you know the calculation method, the benchmark used, the time period and whether fees are already subtracted. Since few active funds hold positive alpha over a decade or more, look for persistent, risk-adjusted alpha that exceeds the fees you would pay for it.