Portfolio and risk · Measuring return and risk
Alpha vs. beta: market return and the return beyond it
How alpha and beta split a stock’s return into a market part and the rest, worked on ten stocks over two years, and why one period’s alpha proves little.
On this page
Beta measures how much of an investment's return moves with the market. Alpha is the part that is left over: the return the investment earned above (or below) what its beta alone would have delivered. Together they split any return into two pieces, the market's and the investment's own, which is why fund managers sell "alpha" and index funds sell cheap "beta".
The split is mechanical, and its result depends heavily on the period. Over the two years to November 17, 2026, Johnson & Johnson's beta to the S&P 500 fund SPY was 0.02 and its alpha 27.5% a year, the best of ten large stocks. Nvidia's beta was 1.83 and its alpha −4.1%. Neither number says the next two years will look the same. This page shows how alpha and beta are measured, works them for ten stocks, and explains what each can and cannot tell you.
The two measures side by side
| Beta | Alpha | |
|---|---|---|
| What it measures | Sensitivity to the market | Return beyond what the market exposure explains |
| Units | A ratio (market = 1.0) | Percent per year |
| Neutral value | 1.0 (moves like the market) | 0 (earned exactly its market-based return) |
| Source | Regression slope | Regression intercept |
| Can you buy it cheaply? | Yes, with an index fund | Not reliably |
| Persistent over time? | Fairly, for broad groups | Rarely, for individual managers or stocks |
| Full lesson | Portfolio beta: how much of the market a portfolio carries | Alpha: the return that market exposure does not explain |
This is Jensen's alpha, introduced by Michael Jensen in 1968 to judge mutual funds. It is the standard meaning; "alpha" used loosely to mean "beat the index" ignores beta and can mislead.
Ten stocks, one market, two years
All figures use daily price returns from November 18, 2024 to November 17, 2026 (500 trading days), with SPY as the market. Over those two years SPY returned 11.7% a year and 3-month Treasury bills about 4.1%.
| Stock | Return a year | Beta | CAPM expected return | Alpha |
|---|---|---|---|---|
| Johnson & Johnson | 34.5% | 0.02 | 4.3% | 27.5% |
| Coca-Cola | 26.6% | −0.04 | 3.8% | 21.5% |
| Exxon Mobil | 18.0% | 0.16 | 5.4% | 14.3% |
| Apple | 24.4% | 1.04 | 12.0% | 13.3% |
| JPMorgan Chase | 16.2% | 0.90 | 11.0% | 6.1% |
| Berkshire Hathaway B | 11.6% | 0.35 | 6.8% | 5.3% |
| Microsoft | 10.4% | 0.94 | 11.2% | 2.3% |
| Tesla | 6.0% | 2.21 | 20.9% | −1.0% |
| Procter & Gamble | 0.3% | 0.10 | 4.9% | −2.7% |
| Nvidia | 5.4% | 1.83 | 17.9% | −4.1% |
The regression alpha and "return minus CAPM expected return" differ slightly because the first is an average of daily differences and the second compares compounded yearly rates; both tell the same story.
What the comparison teaches
Beta sets the bar. A stock with a high beta has to beat a high hurdle to show positive alpha. Tesla's 6.0% a year was not far from Procter & Gamble's 0.3% in alpha terms, because Tesla's beta of 2.21 implied that a market rising 11.7% should have lifted it by about 21%.
Raw return is not alpha. Coca-Cola's 26.6% and Apple's 24.4% look similar, but Apple's beta of 1.04 means about 12 points of its return were simply the market. Coca-Cola's came almost entirely from outside the market.
Two years is a small sample. The standard error of an alpha is roughly the stock's residual volatility divided by the square root of the number of years. For these stocks, with residual volatility of 16% to 42% a year, that is about 11 to 30 percentage points over two years. Every alpha in the table is within two standard errors of zero. A big alpha in one window is consistent with luck, and the same stocks rank very differently in other periods.
Low betas move. Over this window several defensive stocks showed betas near zero. Over longer history, the consumer staples sector fund XLP had a monthly beta of about 0.5 since 1999. A beta measured in an unusual period overstates or understates the hurdle, which changes the alpha.
Alpha and beta in funds
The same split is how fund performance is judged. An index fund delivers beta of about 1 and an alpha close to minus its fee. An active fund promises positive alpha. S&P Dow Jones Indices' SPIVA reports have found for many years that most actively managed US stock funds trail their benchmark over 10 and 15 years, which means negative alpha after fees for most of them (see Index investing: passive vs. active funds and what the data shows).
"Smart beta" funds sit between the two: they buy rule-based exposures, such as value or low volatility, that used to be counted as alpha and are now treated as additional betas, or factors (see Factor models: explaining a portfolio’s returns with several betas).
Which to use when
| Question | Use | Why |
|---|---|---|
| How much will this stock add to my market exposure? | Beta | Sensitivity to the market |
| Did a fund manager add value beyond the market? | Alpha (over many years, after fees) | Return beyond the exposure taken |
| What return should I demand for this risk? | Beta, through CAPM | Sets the hurdle |
| Is a stock's past outperformance skill or market exposure? | Both | Beta explains the market part; alpha is the rest |
| Is a strategy's edge real? | Alpha with its standard error, out of sample | One period's alpha is mostly noise |
How Finabase does it
Each symbol's overview lists the stock's market beta among its key statistics, and the Screener has a Beta column for filtering whole markets by it. The comparison tab draws a stock against SPY or any other benchmark, which is the quickest way to see where a stock's return came from; the link above opens Coca-Cola against SPY. (The chart's BETA study is a different calculation, a short TA-Lib statistic on each bar's highs and lows, not market beta.) Finabase does not publish an alpha figure for stocks or funds: as this page shows, alpha depends on the window, the benchmark and the risk-free rate chosen, and it is easy to over-read. The numbers above were computed by Finabase from daily price returns (no dividends) and 3-month Treasury bill rates.
Frequently asked questions
What is the difference between alpha and beta?
Beta measures how much an investment moves with the market. Alpha measures the return it earned beyond what that market exposure would have produced. Beta is the market's share of the return; alpha is the investment's own.
Is a high alpha good?
A positive alpha over a long period, after fees and on the right benchmark, is what active managers aim for. Over a short period a high alpha is often luck, and the next period frequently reverses it.
Can alpha be negative?
Yes. A stock or fund that earns less than its beta implied has negative alpha, as Nvidia did over these two years (−4.1% a year). Index funds typically have a small negative alpha equal to their fees.
How long a record do I need to judge alpha?
Longer than most people expect. For a fund whose returns stray from its beta by 6% a year, an alpha of 2% a year needs (2 × 6 ÷ 2) squared, or about 36 years of data, before it is two standard errors from zero. For a single stock with 20% a year of its own volatility, the same test would take about 400 years. That is why studies of fund skill use decades of returns and many funds at once, and why a manager's three good years say little.
Do I need alpha to do well?
No. Market returns, the beta part, have historically done most of the work for long-term investors, and they can be bought cheaply through index funds.
Summary
Beta is the market part of a return; alpha is what is left after accounting for it. Over the two years to November 17, 2026, with SPY returning 11.7% a year and T-bills 4.1%, Nvidia's beta of 1.83 set a 17.9% hurdle it missed by 12.5 points, while Johnson & Johnson's near-zero beta turned almost all of its 34.5% return into alpha. Use beta to judge market exposure and set the required return, use alpha over long periods and after fees to judge skill, and treat any short-period alpha as a hypothesis rather than a result.
Check your understanding
4 questions on this lesson. Each answer comes with the reason, and nothing is scored.
T-bills yield 4%, the market returns 12%, and a stock with a beta of 1.5 returns 15%. What is its alpha, approximately?
- +3%
- −1%
- +11%
Show the answer
B. −1% CAPM expected return = 4% + 1.5 × (12% − 4%) = 16%. Actual 15% − expected 16% = −1%. Beating the market by 3 points was not enough for a stock with that much market risk.
Coca-Cola returned 26.6% a year with a beta near zero; Apple returned 24.4% with a beta of 1.04. Why is Coca-Cola’s alpha larger?
- Coca-Cola paid a bigger dividend
- Apple’s return was negative
- About 12 points of Apple’s return are explained by its market exposure; almost none of Coca-Cola’s are
Show the answer
C. About 12 points of Apple’s return are explained by its market exposure; almost none of Coca-Cola’s are Alpha is return beyond what beta predicts. With a beta near zero, the hurdle is roughly the T-bill rate, so nearly all of Coca-Cola’s return counts as alpha.
A stock showed an alpha of +20% a year over two years. What is the most careful reading?
- It will keep beating the market by 20% a year
- It may be luck: with typical stock-specific volatility, two years of data leave an alpha uncertain by 10 to 30 points
- Its beta must be zero
Show the answer
B. It may be luck: with typical stock-specific volatility, two years of data leave an alpha uncertain by 10 to 30 points The standard error of alpha is roughly residual volatility divided by the square root of the years. Short-period alphas are mostly noise.
What does an S&P 500 index fund deliver, in alpha and beta terms?
- Beta of about 1 and alpha of about minus its fee
- Beta of 0 and alpha equal to the market return
- High alpha and low beta
Show the answer
A. Beta of about 1 and alpha of about minus its fee An index fund tracks the market, so its beta is about 1, and it trails the index by roughly its costs, a small negative alpha.
Sources and further reading
- Michael C. Jensen, “The Performance of Mutual Funds in the Period 1945–1964”, Journal of Finance (1968)
- William F. Sharpe, “Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk”, Journal of Finance (1964)
- S&P Dow Jones Indices — SPIVA U.S. Scorecard (active funds against their benchmarks)
- Federal Reserve Board — H.15 Selected Interest Rates, 3-month Treasury bill (FRED series TB3MS)
- Finabase chart data — daily closes 2024-11-18 to 2026-11-17 (price returns, no dividends)
