From July 1963 to August 2026, the US shares that moved most with the market returned 14.1% a year on average. The Capital Asset Pricing Model (CAPM) predicted more, at 15.2%. The shares that moved least returned 10.8%. CAPM predicted less, at 9.1%.
How much a share moves with the market is called its beta. CAPM says beta alone sets the return an investor should expect. This guide explains the model in plain words, then tests it on 63 years of US shares and on 15 funds. It ends with why CAPM is still used to estimate a company's cost of equity, the yearly return its shareholders expect.
What is CAPM? The Capital Asset Pricing Model (CAPM) says an investment's expected return equals a safe interest rate plus its beta times the stock market's extra return over that rate. Beta measures how much the investment tends to move with the market.
On US shares from 1963 to 2026, higher beta did come with higher average returns. But the rise was about half the size CAPM predicts.
On 15 funds from 2004 to 2026, the gold fund had a beta of 0.097, yet it returned 11.3% a year on average. CAPM predicted 2.7%.
CAPM is still a simple way to estimate a company's cost of equity. Eugene Fama and Kenneth French find its answers too high for high-beta shares and too low for low-beta ones.
What is CAPM?
William Sharpe published CAPM in 1964. He received the Nobel Prize in economics in 1990. The model says:
Expected return = risk-free rate + beta × (market return − risk-free rate)
It has three parts:
The risk-free rate is the return on a safe loan to the US government, such as a Treasury bill or a 10-year Treasury bond.
Beta is how much an investment has tended to move when the market moves 1%. A beta of 1.5 means a 1.5% move, on average. What Is Stock Beta? shows how it is measured.
The market return minus the risk-free rate is the equity risk premium: the extra return for owning shares over safe bonds. What Is the Equity Risk Premium? measures it.
In How to Estimate the Cost of Equity, Microsoft's figures were 5.18% + 1.11 × 4.14%. That gives 9.8% a year.
What CAPM predicts
In the model, an investor can spread money across many shares. Each company's own news then cancels out. Only the moves shared with the whole market are left. Beta measures those. So CAPM pays a premium for beta and for nothing else.
That gives two predictions that can be tested:
Returns rise in a straight line with beta. A beta of 0 earns the risk-free rate. A beta of 1 earns the market's return. A beta of 1.5 earns the risk-free rate plus 1.5 times the market's extra return.
Beta is all that matters. Two investments with the same beta should earn the same.
Kenneth French, a finance professor at Dartmouth, publishes free data on US shares. His "Portfolios Formed on Beta" sort shares on the New York, American and Nasdaq exchanges into five groups each June. The sort uses each share's beta over the previous five years. Larger companies count for more within each group.
We took the monthly returns of the five groups from July 1963 to August 2026. We measured each group's beta against the whole US market over that span. One-month Treasury bills returned 4.35% a year on average. The market returned 11.57%, so its extra return was 7.22 points.
Each group's CAPM prediction is 4.35% plus its beta times 7.22 points. CAPM predicts average returns, so we compare it with each group's mean monthly return times 12.
Source: Kenneth R. French Data Library; YX Insights
The lowest-beta group, with a beta of 0.66, returned 10.8% a year against a prediction of 9.1%. The highest-beta group, at 1.50, returned 14.1% against 15.2%. The middle three groups landed close to their predictions.
So higher beta did pay, but less than promised. Each extra unit of beta added 4.0 points a year across the five groups. CAPM predicts 7.2 points. That is about half.
The pattern also changed over time. From July 1963 to December 1990, higher beta did not pay at all. The highest-beta group returned 11.5% a year on average. The lowest-beta group returned 11.9%.
What the classic studies found
Published studies point the same way:
Black, Jensen and Scholes (1972) grouped shares by beta. Returns rose with beta, but along a flatter line than CAPM predicts. Low-beta groups beat their predictions, while high-beta groups fell short.
Fama and French (1992) tested US shares from 1963 to 1990. Once they allowed for beta's link with company size, they found the link between beta and average return was "flat". That held even when beta was the only thing used to explain returns. Company size explained more. So did a company's accounting value compared with its market value.
Frazzini and Pedersen (2014) found that high beta came with low alpha, the return above CAPM's prediction. That held in US shares, 20 other stock markets, Treasury bonds, corporate bonds and futures.
Our figures since 1963 fit this record. The lowest-beta group beat CAPM by 1.7 points a year. The highest-beta group fell short by 1.1 points.
Testing CAPM on 15 funds
We ran the same test on 15 funds used across our guides. They are SPY (the S&P 500 exchange-traded fund, or ETF), QQQ (the Nasdaq-100 ETF), IWM (the Russell 2000 ETF) and EFA (the developed-markets ETF outside the US and Canada). The rest are nine US sector funds, GLD (the gold ETF) and TLT (the long-term US Treasury bond ETF).
The window runs from December 2004, after GLD launched, to September 2026. Each fund's beta is measured against SPY on 262 monthly returns. Three-month Treasury bills returned 1.78% a year on average. SPY returned 11.53%.
French's file comes with a one-month bill rate, so the share test uses that. For the funds, we use the three-month rate published on FRED, the St. Louis Fed's database.
Source: YX Insights
The dashed line is what CAPM predicts. The orange line is the best fit through the 15 funds. Each extra unit of beta came with 3.8 points a year more return, against 9.8 points in CAPM. The line through the funds is less than half as steep.
Fund | Beta to SPY | Average yearly return | CAPM prediction (from unrounded inputs) | Gap (points) |
|---|---|---|---|---|
TLT (long-term US Treasury bonds) | −0.075 | 3.7% | 1.0% | 2.7 |
GLD (gold) | 0.097 | 11.3% | 2.7% | 8.6 |
XLU (utilities) | 0.462 | 9.4% | 6.3% | 3.1 |
XLP (consumer staples) | 0.566 | 9.2% | 7.3% | 1.9 |
XLV (health care) | 0.690 | 10.8% | 8.5% | 2.3 |
EFA (developed markets outside the US and Canada) | 0.965 | 7.5% | 11.2% | −3.7 |
SPY (S&P 500) | 1.000 | 11.5% | 11.5% | 0.0 |
XLE (energy) | 1.058 | 11.7% | 12.1% | −0.4 |
QQQ (Nasdaq-100) | 1.128 | 16.0% | 12.8% | 3.2 |
XLI (industrials) | 1.130 | 11.5% | 12.8% | −1.3 |
XLK (technology) | 1.131 | 16.6% | 12.8% | 3.8 |
XLB (materials) | 1.135 | 9.7% | 12.9% | −3.2 |
XLY (consumer discretionary) | 1.148 | 11.6% | 13.0% | −1.4 |
IWM (Russell 2000) | 1.188 | 10.1% | 13.4% | −3.3 |
XLF (financials) | 1.210 | 8.0% | 13.6% | −5.6 |
Source: YX Insights
The biggest misses sit at both ends:
GLD had a beta of 0.097, yet it returned 11.3% a year against a prediction of 2.7%.
XLF (the financials sector ETF) had the highest beta, at 1.21. It returned 8.0% against 13.6%.
XLK (the technology sector ETF) had a beta of 1.13. It returned 16.6% against 12.8%.
Beta also missed most of the risk in some funds. Over the same window, TLT swung by 13.6% a year, as measured by its volatility, yet its beta was −0.075. Volatility is covered in What Is Volatility?
The market explained less than 1% of the monthly moves in TLT and GLD. This is one 22-year window and 15 funds, so one strong run can shape the result.
Why CAPM is still used
CAPM needs only three inputs. It gives a number for any listed company. It also ties the return shareholders expect to a risk that can be measured.
In a 2001 survey of 392 chief financial officers, published in the Journal of Financial Economics, John Graham and Campbell Harvey found that large firms rely heavily on CAPM. The CAPM cost of equity then feeds a company's Weighted Average Cost of Capital (WACC), as in What Is WACC?
How to read a CAPM estimate
Fama and French spell out the cost of the flatter line. In their 2004 review, CAPM cost of capital estimates "for high-beta stocks are too high" and those "for low-beta stocks are too low", compared with past returns.
Treat it as a central estimate. Test how much a valuation moves when the cost of equity changes by a point.
Check the beta. The window and the return interval change it, as What Is Stock Beta? shows.
Expect misses at the ends. Since 1963, very low betas have beaten CAPM, while very high betas have fallen short.
Compare methods. The cost of equity guide sets CAPM against two other ways to estimate it.
CAPM gets the direction right but not the size. Since 1963, higher-beta shares earned more. But the extra return was only about half what the model predicts. So treat a CAPM cost of equity as a central estimate with a range around it.
Learn more with YX Insights
This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.
The same approach runs through everything else we publish:
Systematic Portfolio: ready-made portfolios for Macro & Megacaps and for Commodities. We publish the holdings and every change.
Multi-model Signals: a daily long-or-flat call on every name we cover. Each call shows how strongly our four models agree.
Research: company deep dives, macro commentary and essays on how we test.
Good places to start on the website:
Academy: more explainers like this one.
Start here: get our free weekly letter by email.
Macro & Megacaps Systematic Portfolio and Commodities: the two portfolios in detail.
Track record: live results since 3 August 2026, plus the ten-year backtest.
Upgrade: get the Systematic Portfolio, Multi-model Signals and research in full.
Common questions about CAPM
What is the CAPM formula?
The CAPM formula is: expected return = risk-free rate + beta × (market return − risk-free rate). The bracket is the equity risk premium. Take Microsoft in September 2026: a 10-year Treasury yield of 5.18%, a beta of 1.11 and a premium of 4.14%. CAPM then gives a cost of equity of 9.8% a year.
Is CAPM still used?
Yes. CAPM is still a simple way to estimate a company's cost of equity. In a 2001 survey of 392 chief financial officers, John Graham and Campbell Harvey found that large firms rely heavily on it. Its main weakness shows in the data: high-beta shares have earned less than it predicts, while low-beta shares have earned more.
What are the main problems with CAPM?
The main problem is that returns have risen with beta far less than CAPM predicts. On US shares from 1963 to 2026, each extra unit of beta added 4.0 points a year, against 7.2 points in the model. Beta also changes with the window used to measure it. Other traits, such as company size, explain returns too.
What does beta mean in CAPM?
Beta measures how much an investment has tended to move with the stock market. A beta of 1 means it moved in line with the market, on average. A beta of 1.5 means it moved 1.5% for each 1% market move. In CAPM, a higher beta means a higher expected return.
Does a higher beta mean higher returns?
Yes on average, but by less than CAPM predicts. From July 1963 to August 2026, the highest-beta group of US shares returned 14.1% a year, against 10.8% for the lowest-beta group. CAPM predicted a wider gap: 15.2% against 9.1%. From 1963 to 1990, the highest-beta group earned slightly less than the lowest-beta group.
DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.