TL;DR. Beta tells you how much a stock or portfolio moves when the market moves. CAPM — the Capital Asset Pricing Model — turns that beta into a required return: the rate of return investors should demand for taking that much risk. Alpha is the part of a manager’s return that CAPM cannot explain — the skill premium, if it is real. These three numbers are the working vocabulary of modern portfolio theory, and they show up on every fund factsheet, every analyst report, and every cost-of-capital calculation in corporate finance.
Where these ideas came from
Modern portfolio theory began with Harry Markowitz’s 1952 paper on diversification — the insight that an asset’s risk should be measured by what it contributes to a portfolio, not in isolation. A decade later, William Sharpe extended that into the Capital Asset Pricing Model in his 1962 paper, sharing the 1990 Nobel Prize in Economic Sciences with Markowitz and Merton Miller for the work. CAPM is now the most-taught asset-pricing model on Earth and the default discount-rate tool inside investment-banking valuations.
Beta: how much a stock dances with the market
Beta is a single number that answers one question: when the broad market moves 1%, how much does this stock typically move? By construction the market portfolio itself has a beta of 1.0. Cash has a beta of 0. Anything above 1 is more volatile than the market; anything below 1 is less volatile.
The mechanical definition is:
βi = Cov(Ri, Rm) / Var(Rm)
Which is just the slope of the regression of stock i’s returns on the market’s returns — usually estimated with two to five years of weekly or monthly data against an index like the S&P 500. A beta of 1.3 means: historically, a 1% S&P move produced about a 1.3% move in this stock; a beta of 0.5 means about half the move.
Industry differences are large and persistent. Cyclical, capital-intensive businesses have high betas. Regulated utilities have low ones. Here is how it actually looks across a dozen U.S. sectors using the same dataset academics and bankers use:
| Sector | Firms | Unlevered Beta |
|---|---|---|
| Software (Internet) | — | 1.55 |
| Semiconductor | — | 1.49 |
| Auto & Truck | 33 | 1.27 |
| Drugs (Pharmaceutical) | 228 | 0.89 |
| Retail (General) | 23 | 0.76 |
| Tobacco | 10 | 0.68 |
| Food Processing | 78 | 0.46 |
| Telecom Services | 39 | 0.37 |
| Bank (Money Center) | 15 | 0.34 |
| Utility (General) | 14 | 0.15 |
The two-decade gap between an internet-software beta of 1.55 and a utility beta of 0.15 is the entire reason CAPM exists: those two businesses face very different cost of capital, and a one-size-fits-all hurdle rate will systematically over-invest in one and under-invest in the other.
Levered vs unlevered beta
The number you read in a finance textbook is usually the levered (equity) beta — what you actually observe when you regress stock returns on the market. The unlevered (asset) beta strips out the impact of how much debt the company carries. Two companies in the same industry can have very different equity betas just because one is more leveraged. Stripping that out gives you the underlying business risk, which is what you need when you are valuing a private company or building a comparables set.
CAPM: turning beta into a required return
CAPM’s claim is that, in equilibrium, the expected return on any asset is the risk-free rate plus a risk premium proportional to its beta:
E(Ri) = Rf + βi × (E(Rm) − Rf)
In English: take what you can earn on a Treasury bill, then add beta times the equity risk premium — the historical extra return stocks have earned over bonds. That second term is the price the market charges you for taking equity risk, and beta scales it up or down for the specific asset.
Plug in real numbers as of late May 2026. According to the Federal Reserve’s H.15 release, the 3-month Treasury-bill rate was 3.59% on May 19, 2026, and the 10-year Treasury yielded 4.67%. Damodaran’s long-run dataset puts the geometric-mean equity risk premium over T-bonds at roughly 4.9 percentage points using 1928–2025 data.
So for a stock with a beta of 1.2, using the 10-year as the risk-free proxy:
E(R) = 4.67% + 1.2 × 4.9% = 4.67% + 5.88% = 10.55%
For a utility with a beta of 0.30: E(R) = 4.67% + 0.30 × 4.9% = 6.14%. Same risk-free rate, same equity risk premium — very different required return, because the business has very different exposure to the market.
That is also why every CFO uses CAPM: the equity cost figure flows directly into the weighted-average cost of capital that prices every project, every acquisition, and every internal capital allocation decision.
The Security Market Line
If you graph CAPM in beta–return space, you get a straight line: the Security Market Line. Every asset, in CAPM’s view, should sit on that line. Anything above it is offering too much return for its risk — a buy. Anything below it is too little — a sell.
Alpha: what is left over
If you actually measure a portfolio’s return and subtract the CAPM-required return for its beta, what is left is called alpha — the part of performance not explained by market risk. Jensen’s alpha, the version named for Michael Jensen’s 1968 paper, is defined as:
α = Rportfolio − [Rf + β × (Rm − Rf)]
Positive alpha means the manager earned more than what their risk exposure would predict. Negative alpha means they earned less. A fund that returned 12% in a year where the CAPM-implied return for its beta was 10% delivered 2 percentage points of alpha — assuming the beta and the model are right.
The catch: most active equity managers do not generate persistent positive alpha after fees. S&P Dow Jones’ SPIVA scorecards have shown for two decades that the majority of large-cap U.S. funds fail to beat their benchmark over five- and ten-year horizons. That is the empirical workhorse behind the index-fund movement.
Where the model breaks down
CAPM is the most-taught model in finance, but it is also the most-attacked. The three big critiques you should know:
- Beta does not capture everything. Fama and French showed in 1992 and 1993 that company size and value (book-to-market) explain a large chunk of cross-sectional returns that beta alone misses. The Fama-French three-factor model and later five-factor model are the academic response.
- Betas are unstable. A regression on the last five years of returns assumes the future will look like the past. Companies change capital structure, business mix, and competitive position; their true beta drifts with them.
- The market portfolio is unobservable. CAPM says you should regress against the true market — every risky asset on Earth, weighted by value. In practice, people use the S&P 500, which Richard Roll’s 1977 critique pointed out is a real problem if you take the model literally.
None of this makes CAPM useless. It makes it a first approximation — a clean lens that captures most of what matters in equity pricing, while leaving room for factor models, fundamental analysis, and judgement to fill the rest.
The data in pictures: stocks, bonds, bills since 1928
The reason equity beta gets compensated at all is that, over time, stocks have delivered enough extra return over bonds and bills to make the volatility worth it. Damodaran’s 1928–2025 series tells the story most cleanly:
That ~5-percentage-point gap between the green and blue bars is the long-run equity risk premium — the price the market has historically paid investors for putting up with equity volatility. It is the number that, multiplied by beta in the CAPM formula, gives you a stock-specific required return.
How to actually use these numbers
For an individual investor, the most useful thing about beta, alpha, and CAPM is what they help you avoid. Three concrete uses:
- Sanity-check a fund’s claimed returns. If a fund averaged 15% last year but its beta is 1.6 and the market did 9%, its CAPM expectation was roughly the risk-free rate plus 1.6 × the equity risk premium — pretty close to that 15%. The manager mostly delivered beta, not alpha.
- Right-size your equity exposure. If your goal portfolio beta is 0.7, you can blend a S&P 500 ETF with cash or short-duration Treasuries to land there. The math is linear: 70% market plus 30% T-bills gets you a portfolio beta of 0.7.
- Build a hurdle rate for your own decisions. Buying a rental property, a small business, or a private investment? You can run a CAPM-style calculation using a comparable public-company beta to get a rough required return and refuse anything that does not clear it.
What to learn next
If beta and CAPM clicked for you, the natural next steps are the Fama-French and Carhart factor models (size, value, momentum), and the Sharpe ratio — a different and arguably more honest measure of risk-adjusted return that divides excess return by total volatility instead of beta. For the corporate-finance side, the weighted-average cost of capital (WACC) is where CAPM’s equity-cost estimate gets blended with the after-tax cost of debt to discount real-world projects.
Sources
- Nobel Prize — William Sharpe (1990) facts and CAPM citation
- Nobel Prize — Harry Markowitz (1990) facts
- Federal Reserve H.15 Selected Interest Rates — T-bill and Treasury yields, 19 May 2026
- Aswath Damodaran — Unlevered beta by sector (January 2026)
- Aswath Damodaran — Historical Returns on Stocks, Bonds and Bills 1928–2025
- Fama-French three-factor model (1992–93 critique of CAPM)
- S&P Dow Jones SPIVA scorecards on active manager performance
Disclosure: This article was produced with AI assistance and reviewed before publication. It is for informational purposes only and is not investment advice.